An Investigation Into Factors that, Affect Monoclonal Antibody ...

An Investigation Into Factors that, Affect Monoclonal Antibody
Production by Hybridomas in Culture.
by
P. M. Hayter
Department of Microbiology
University of Surrey
Guildford
Surrey.
Submitted for the degree of Doctor of Philosophy.
September, 1989.

ABSTRACT
The purpose of this investigation was to determine the
effects of the culture environment on antibody production by
hybridomas. Several observations made during" this Investigation
suggest that antibody production by the 321 hybridoma was
favoured at low growth rates.
1) Antibody production by the 321 hybridoma continued throughout
the growth and decline phases of batch cultures but the specific
rate of antibody production was highest when cell growth had
ceased.
2) In proliferating cells, the antibody production rate was higher
at pH 6.8 or with 1% newborn calf serum than at pH 7.2 and with
1094 serum, conditions which were more favourable for cell growth.
3) Antibody production rates were higher in cultures where cell
division had been arrested by excess thymidine.
4) Specific antibody production rates and the yield of antibody from
glutamine were higher at low growth rates In chemostat culture.
Studies on the metabolism of the 321 hybrldoma provided
evidence for an increased demand on metabolite and energy pools
at high growth rates suggesting that the cell's metabolic capacity
may be diverted towards cell growth at the expense of antibody
synthesis.
Alternatively, antibody production may be cell cycle-related.
In synchronised cultures of the 321 hybridoma there was a
reproducible decline in the antibody production rate during mitosis

and evidence of a further decline during the S phase. Higher rates
of antibody synthesis during the Gx and G2 phases suggest that an
increase in the proportion of cells residing in one or other of those
phases would result in an increase in the rate of antibody
production. Observations made on the effect of pH on the cell cycle
were consistent with the hypothesis that unfavourable
environmental conditions increase the duration of the Gi phase and
hence the proportion of cells residing in that phase.

ACKNOWLEDGEMENTS
I would like to thank my supervisors, Professor R. E. Spier
and Dr N. F. Kirkbyfor their interest and enthusiasm throughout
the course of this work.
I would like to acknowledge the excellent technical assistance
of Di Simpson for her help in tissue culture, Karel Newman for
designing and building the computer interface and Terry Fieldus for
building the continuous filter device.
My friends and colleagues at Surrey during this work, Jane
Thorpe, Steve Musgrave, Bob Wilson, Tim Clayton, Jimmy Modha,
Andrew Murdin, Peter Whiteside, Ric Jordan, Teresa Huck and Janet
Bunker deserve special thanks for their help, encouragement, advice
and above all, friendship.
The cell line used in this investigation was a gift from Dr
A. Wright, Central-Toxicology Laboratory, ICI and I would also like
to thank Stuart Walsh in the same laboratory for helpful
discussions on the purification of the 321 °. antibody.
This work was funded by an Science and Engineering
Research Council studentship.

ABBREVIATIONS
APR antibody production rate
ATP adenosine triphosphate
BSA bovine serum albumin
cpm counts per minute
D dilution rate
DEAE diethylaminoethyl
DMSO dimethyl sulphoxide
ELISA enzyme-linked immunosorbent assay
g gravity constant
Gi phase period following mitosis and preceding DNA synthesis
G2 phase period following DNA synthesis and preceding mitosis
h hours
HPLC high performance liquid chromatography
HPRT hypoxanthine phosphoribosyl transferase
IgG immunoglobulin G
KL a volumetric mass transfer coefficient
M phase period of mitosis
NAD nicotinamide adenine dinucleotide
NADH reduced form of NAD
PAGE polyacrylamide gel electrophoresis
PBS phosphate buffered saline
S phase period of DNA synthesis
SDS sodium dodecyl sulphate

TCA trichloracetic acid
Tris tris(hydroxymethyl)aminomethane
Tween 20 non ionic detergent
11 specific growth rate

Chapter 1. Introduction.
Contents.
1.1 Animal cell culture. 1
1.2 The B-cell hybridoma.
--
1.3 Applications for monoclonal antibodies. 9
1.4 Production of monoclonal antibodies. 12
1.4.1 Ascites. 12
1.4.2 In vitro production methods. 13
1.4.2.1 Immobilised cell systems. 14
1.4.3 Assessment of bloreactor types. 18
1.5 Hybridoma growth and production kinetics.
1.6 The effect of the environment on cell growth and
6
, 20
antibody production. 23
1.6.1 Carbohydrates and cell growth. 23
1.6.2 The effect of glucose and lactate on
antibody production. 25
1.6.3 Amino acid metabolism.
1.6.4 The effect of glutamine on antibody
production. 31
1.6.5 The effect of ammonia on cell growth
and antibody production. 32
1.6.6 Oxygen requirements by animal cells
in culture.
26
33

1.6.7 The effect of oxygen on antibody
production.
1.6.8 The effect of pH on cell growth and
antibody production. 36
1.6.9 The role of serum in cell growth. 37
1.6.10 The effect of serum on cell growth and
antibody production. 40
1.7 Cell cycle-related changes in protein expression. 41
1.8 Summary and objectives. 44
Chapter 2. Materials and methods.
2.1 Cell culture.
2.1.1 Cell line. 46
2.1.2 Media. 46
2.1.3 Cell culture. 47
2.1.4 Cryopreservation. 47
2.2 Assay techniques. 48
2.2.1 Glucose determination. 48
2.2.2 Lactate determination. 49
2.2.3 Ammonia determination. 50
2.2.4 Glutamine determination. 51
2.2.5 Antibody determination. 52
2.3 Purification and characterisation of the 321
antibody standard. 54
35
46

2.3.1 Preparation of paraquat-AH Sepharose
affinity column. 54
2.3.2 Affinity purification of the 321 antibody. 56
2.3.3 Determination of the protein content of
pooled fractions. 56
2.3.4 Determination of the antibody purity by
SDS-polyacrylamide gel electrophoresis. 5;
2.3.5 Assessment of antibody purity by high
performance liquid chromatography. 62
Chapter 3. The automated assay system.
3.1 Requirements for an automated assay system. 65
3.1.1 Methods for the monitoring of product
formation in fermenters. 66
3.2 Immunosorbent preparation. 69
3.2.1 Preparation of the IgG fraction from
sheep serum. 69
3.2.2 Coupling antibody to Sepharose CL4B. 70
3.2.3 Assessment of the Immunosorbent in the
ELISA. 74
3.2.4 Continuous flow ELISA. 77
3.3 The automated assay. 82
3.3.1 Computer operation of the assay. 85

3.3.2 Sequence of events during the automated
assay.
3.4 Optimisation of the automated assay. 87
3.4.1 Determination of the optimum labelled
antibody dilution. 87
3.4.2 Buffer composition. 87
3.5 Assay performance. 98
3.6 The use of the automated ELISA for monitoring
bioreactors. 101
3.6.1 Column compression. 101
3.6.2 On-line sampling of the bioreactor. 104
3.7 Conclusions. 107
Chapter 4. Hybridoma growth and antibody production kinetics in
batch culture.
4.1 Introduction. 109
4.2 Materials and methods. 110
4.3 Results., 110
4.3.1 Hybridoma growth and antibody production in,
static flasks. 110
4.3.2 Hybridoma growth and antibody production
in stirred batch culture. 113
4.3.3 The effect of glutamine on cell, growth
and antibody production.. 118
86

4.3.4 The effect of pH on cell growth and
antibody production. 124
4.3.5 The effect of serum on cell growth and
antibody production. 125
4.3.6 The effect of excess thymidine on cell
growth and antibody production. 132
4.3.7 The effect of glutamine concentration on
thymidine-blocked cells. 135
4.3.8 The effect of pH on thymidine-blocked
cells.
4.3.9 The effect of serum concentration on
137
thymidine-blocked cells. 141
4.4 Discussion. 141
Chapter 5. Continuous culture.
5.1 Introduction. 151
5.2 Materials and methods. 153
5.2.1 Apparatus for the continuous culture of
hybridoma cells. 153
5.2.2 Continuous culture. 154
5.2.3 Metabolic quotients. 156
5.3 Results. 158
5.3.1 Establishment of steady state growth. 158
5.3.2 The effect of the specific growth rate on
the metabolism of the 321 hybridoma. 161

5.3.2.1
5.3.2.2
5.3.2.3
5.3.2.4
5.3.3 Hybrido
5.4 Discussion.
Cell numbers and viability. 161
Glutamine and ammonia. 167
Glucose and lactate. 168
Antibody. 168
ma stability.
Chanter 6. Synchronised culture.
6.1 Introduction.
168
170
180
6.1.1 Physical selection methods. 182
6.1.2 Growth arrest. 183
6.2 Material s and methods.
6.3 Results.
184
6.2.1 Synchronisation of the 321 hybridoma. 184
6.2.2 Radiolabelling of proteins. 185
6.2.3 Radiolabelling of DNA. 185
6.2.4 SDS-polyacrylamide gel electrophoresis. 186
6.3.1 Determination of the S-phase 186
6.3.2 Cell synchronisation. 187
6.3.3 The incorporation of 35 S-methionine into
antibody.
194
6.3.4 The effect of pH on the cell cycle. 197
6.4 Discussion.
6.5 Conclusions.
201
209

Chapter 7. Summary and Recommendations.
7.1 Summary. 211
7.2 Recommendations. 218
References. 221
Appendices.
Appendix I. Solenoid interface. 251
Appendix II. Automated ELISA control program. 252
Appendix III. Cross-flow filter. 258
Appendix IV. The determination of the volumetric mass
transfer coefficient (K,. a) for oxygen in
the Gallenkamp fermenter. 259

1.1 Animal cell culture.
Chapter 1
Introduction
The earliest reports of explanted tissue being maintained and
studied in vitro date back to the late 19th century when Wilhelm Roux
(1885) maintained chick embryo tissue for several days in warm saline
and Arnold (1887) studied the movements of isolated frog leukocytes
maintained in lymph-soaked elder pith over a period of 4-5 days.
However the work that stimulated most interest in tissue culture was
a study on growth of nerve cells in lymph clots carried out by
Harrison in 1907. Subsequent work by Carrel and colleagues (Carrel
and Burrows, 1911; Burrows, 1912; Lewis and Lewis, 1912) led to
refinements in salt and amino acid composition in culture media
containing plasma or clotted lymph as growth promoting supplements.
The first report of serial propagation followed in 1916 when Rous and
Jones dispersed chick embryo cells from tissue and grew them in a
plasma clot. They were able to redisperse the cells using trypsin and
there was subsequent regrowth. These early studies demonstrated that
the structural organisation and physiology of tissues could be
investigated by in vitro culture techniques and in the next decade
studies on cell metabolism established that the energy requirements of
cell were provided by glucose (Lewis, 1922; Warburg, 1930) and/or the
deamination of amino acids (Holmes and Watchorn, 1927; Warburg and
Kubowitz, 1927). The quantitative oxygen requirements of cells were
also examined at this time (Warburg and Kubowitz, 1927).
In the 1930s and 1940s there were many studies concerned with
1

the identification of the essential amino acid, vitamin and inorganic
components in plasma, lymph or serum which were the principle media
for tissue culture at that time. This resulted in the development of
media with defined components (Fisher et al., 1948; White, 1949;
Morgan et al., 1950; Earle et al., 1951) and culminated in Eagles
painstaking determination of the essential amino acid and vitamin
requirements for several mammalian cell lines (Eagle, 1955a; Eagle,
1955b). Modifications of these early media are still extensively used
today.
Since those initial studies tissue culture has proved to be a
valuable tool for the study of animal cell physiology. Following the
discovery by Enders in 1949 that poliomyelitis virus could be grown In
primary cell cultures, cell culture has also been used in virus
propagation and vaccine production. Indeed the large scale use of
animal cell cultures for the production of biological material was
largely confined to virus vaccine production until relatively recently
despite the obvious therapeutic potential of many mammalian proteins
which have been discovered In tissues. However the production of large
quantities of proteins from normal tissues Is often limited by the
availability of the appropriate tissue. the low concentrations present in
tissues and loss of protein during purification. The large scale
production of proteins from cultured normal cells is limited by the finite
lifespan of normal cells in vitro. It was therefore considered that the
future of large scale production of human proteins lay with genetically
engineered bacteria and while biologically active proteins such as
insulin (Johnson, 1983) and human growth hormone (Hsuing et al., 1986)
have been produced in large quantities from Escherichla cola, there are
many problems associated with the expression of mammalian proteins in
2

acteria. For example bacteria are unable to carry out post translational
modifications such as glycosylation, phosphorylation and signal peptide
cleavage which may affect the activity or solubility of the product. The
proteins may also be Incorrectly folded or present in a denatured form
(Williams et al., 1982; Marston, 1986; Kenealy et al., 1987). While lower
eukaryotes such as yeast can glycosylate and secrete proteins, the
glycosylation pattern and protein folding may be Incorrect (Ballou, 1982;
Kingsman et al., 1987) and low product activities have been reported
(Wood et al., 1985). It has therefore become apparent that while a
number of biologically active proteins have been produced in yeast and
bacteria (Berman and Lasky, 1985; Kingsman et al., 1987). proteins such
as erythropoietin, which rely on correct post translational modification
for their activity (Dube et al., '1988), may only be effectively expressed
in mammalian cells. Furthermore the potential for large scale protein
production by animal cells has been advanced by a number of recent
developments in mammalian cell technology.
The first of these was the development of the hybridoma by
Kohler and Milstein in 1975. The potential of monoclonal antibodies as
therapeutic and diagnostic agents was quickly realised and led to the
development of large scale culture processes to provide. sufficient
material for clinical -testing. Secondly the finding that a number of
continuous cell lines constitutively express useful therapeutic proteins
has led to the development of large scale culture processes for the
production of proteins such as tPA from human melanoma cells (Kluft et
al., 1983) and a-interferon from Namalwa cells (Phillips et al., 1985;
Lazar et al., 1987). More recently the development of techniques for the
expression of recombinant human protein in -mammalian cell lines
(Ringold et al., 1981; Lubiniecki, 1987; Bebbington and Hentschel, 1987)
3

has greatly increased the potential for the production of proteins from
mammalian cell cultures.
The main concern with the use of continuous cell lines is the
perceived risk of the transfer of material with tumorigenic potential to
the recipient since some transforming viral genes and activated
oncogenes can transform cells in. vitro (Bishop. 1983; Diderholm et al..
1965) and cause tumours in vivo (Fung et al.. 1983; Israel et al.. 1979).
However improved techniques for the detection of residual nucleic acids
(Van Wezel et al.. 1982) and favourable assessments of the risks
involved (Petricianni. 1987; Ramabhadran. 1987) have led to an
increasing acceptance of products derived from continuous cell lines for
human therapy. The market value of such products is high and animal
cell products also have potential applications In research, diagnostics
and biochemical purification. Comprehensive reviews of current and
potential applications of mammalian cell technology may be found in the
reports of Spier and Horaud (1985), Lubiniecki (1987), Mizrahl (1988)
and Spier (1988) and some examples of these applications are presented
in table 1.1.
This investigation however concentrates on the production of
monoclonal antibodies by hybridoma cell cultures. This system was
chosen for study for several reasons.
1) Monoclonal antibodies have potential in a number of fields
including therapeutic medicine. There is therefore a requirement for the
large scale production of antibodies by hybridomas. The design and
optimisation of such processes will be facilitated by an analysis of the
factors which influence antibody production by hybridomas.
ii) Hybridomas grow readily in homogeneous suspension culture which
is amenable to conventional fermentation techniques.
4

Table 1.1. Current and potential products fron cultured animal cells.
Viral vaccines Spier 1983
Recombinant
viral antigens
hepatitis B surface antigen
Spier and Horaud, 1985
Mizrahi, 1988
Crowley et a1., 1983
herpes simplex g1ya)protein D Berman et al. 1985
Immunoregulatory proteins
interferans, interleukins
Monoclonal antibodies
Polypeptide growth factors
Enzymes
Hormones
tPA
factor ü
factor VIII
Shoham, 1983
Phillips et al., 1985
Klein et al., 1983
Kris et al., 1985
Kadouri and Bohak, 1985
Kaufman et al. 1985
Anson et al., 1985
Toole et al. 1984
Wood et al. 1984
erythropoietin Jacobs et al. 1985
Viral insecticides
Katinger and Bliem, 1983
baculoviruses Miltenberger and Krieg, 1984
71iatr specific antigens
carcinoembryic antigen
Animal cells as products
reconstitution of skin
5
Kim et a1., 1985
Reuveny et a1., 1987
Green et al. 1979
Bell et al. 1983

ill) Hybridomas secrete a product that is readily detected and
characterised by existing assay techniques.
1.2 The B-cell hybridoma.
The B-cell hybridoma is the result of a cell -fusion between a
B-lymphocyte which secretes antibody and a myeloma cell line which
confers immortality to the resulting fusion product. In accordance with
Burnet's clonal selection theory for antibody production by
B-lymphocytes (Burnet, 1957). It is possible to select a single hybridoma
clone producing antibodies of uniform specificity. The hybridisation
technique was developed by Kohler and Milstein in 1975 and has since
become a widely used technique in research. Detailed descriptions of the
technique may be found in several reviews and books (Milstein, 1980;
Fazekas de St Groth and Scheidegger, 1980; Goding, 1983; Campbell,
1984; Epstein and Epstein, 1986) therefore only the basic principles will
be outlined below (see figure 1.1).
To produce a hybridoma secreting an antibody of the desired
specificity an animal (usually a mouse) is Immunised with the antigen
of interest although in vitro immunisation techniques are also used
(James and Bell, 1987; Borrebaeck. 1987). B-lymphocytes, a proportion
of which will be secreting antibody to the antigen, are harvested from
the spleen, and fused with a cultured myeloma cell line. Kohler and
Milstein used Sendai virus to promote cell fusion but the use of
polyethylene glycol (Davidson et al., 1976) is more convenient and Is
more commonly used today. Improved fusion efficiencies have been
claimed for electrofusion
(Zimmerman et al., 1985).
6

Figure 1.1. The principles of hybridoma production
EýS:
Cultured myeloma
Mouse immunised
with antigen
a
B-lyrrphocytes harvested
from spleen
HPRT
- Cell fusion
oooo0
0ö 0 08
Hybridomas selected
with HAT medium
(promoted by PEG)
Screen for cells producing
desired antibody
Select by cloning
7

The fusion partner is a myeloma cell line which is deficient in the
enzyme hypoxanthine phosphoribosyl transferase (HPRT). This enzyme
is part of a salvage pathway which utilises hypoxanthine in the absence
of de novo purine synthesis. Cells lacking HPRT are unable to
synthesize purines from hypoxanthine and will die in the presence of
aminopterin (a folate antagonist which blocks the de novo synthesis of
purines and pyrimidines). Normal B-cells which possess both the
enzymes thymidine kinase (TK) and HPRT can synthesize both purines
from hypoxanthine and pyrimidines from thymidine but are unable to
proliferate in vitro. Successful B-cell myeloma fusions may therefore be
selected by their ability to proliferate in HAT medium which contains
hypoxanthine, aminopterin and thymidine (Littlefield, 1964).
It is then possible to screen the hybridomas for secretion of
the desired antibody by immunoassay and subsequent selection and
cloning will result in a single hybridoma clone producing monoclonal
antibody.
The monoclonal antibody has a number of advantages over more
conventional polyvalent antisera.
I) The antigen binding site Is uniform in its specificity and affinity
whereas polyvalent antisera contain a population of antibodies from
several B cell clones which have a variety of antigen affinities and
specificities.
ii) It is possible to use relatively
Impure antigens In the immunisation
of animals for monoclonal antibody preparation because cells producing
the desired antibody can be selected after cloning. The production of
a polyvalent antiserum requires that the antigen is of the highest
purity to ensure that the greatest proportion of antibodies are directed
8

towards the desired determinants.
ill) The monoclonal antibody, which can be produced in unlimited
quantities, is highly consistent in its properties whereas the quality of
polyvalent antiserum varies from animal to animal and between
bleedings of the same animal.
1.3 Applications for monoclonal antibodies.
The main use of monoclonal antibodies to date has been in the
fields of research and diagnostics. The increasing use of
diagnostic and assay kits in medicine has resulted in a demand for
highly specific and reproducible reagents. The monoclonal antibody
fulfils both these criteria. Furthermore the unique specificity of the
monoclonal antibody has allowed research to progress into areas which
would prove too rigorous for conventional antisera. Such areas include
the identification of tumour-specific antigens (Schlom and Weeks,
1985), the delineation of closely related cell types such as lymphocytes
(Van Wauwe et al., 1981; Talle et al., 1983), cell receptor studies
(Goldfine et al., 1985) and studies on the antigenic variation of viruses
(Wiktor and Koprowski, 1978; Gerhard et al.. 1981). The application of
monoclonal antibodies in these fields has been reviewed extensively
(see for example Epstein et al., 1986) and will not be considered further.
Another growing area is the use of monoclonal antibodies for the
purification of valuable biologically active substances by
immunoaffinity chromatography. This enables the highly specific
separation of a product which may be present In low quantities and
heavily contaminated by other substances. Industrial processes have
been described for the separation of hepatitis B surface antigen (as
9

eported by Spier, 1983) and interferon (Secher and Burke, 1980).
The potential of monoclonal antibodies In therapeutic medicine has
attracted a considerable amount of attention. The use of monoclonal
antibodies for passive immunisation is attractive because, at present,
antibodies have to be fractionated from immune sera from humans or
animals with the risk of transfer of infectious agents. There are now
human monoclonal antibodies to bacterial, viral and parasite antigens
(James and Bell, 1987) with obvious potential in the treatment of
resistant or highly virulent organisms. The potential for such treatment
has already been demonstrated in the veterinary field by the use of
monoclonal antibodies against E. cola K99 antigen to prevent diarrhoea
in neonatal calves (Sadowski, 1982).
Target-directed drug therapy has also received considerable
attention in recent years due primarily to its potential for the
treatment of cancer. Monoclonal antibodies raised against tumour
associated antigens may be conjugated to toxins such as abrin or ricin
which will then selectively destroy cells bearing those antigens (Youle
and Neville, 1980; Gilliland et al., 1980; Lord et al.. 1985; Vitetta et
al.. 1987). This technique may be especially useful in autologous bone
marrow transplants for the removal of tumour cells (Krolick et a!.,
1982; Bast et a!.. 1983) or in allogeneic transplantation to remove
T-cells from donor marrow so reducing the incidence of graft versus host
disease (Vallera eta!.. 1982). Successful treatment of acute graft versus
host disease In humans by the use of ricin conjugated antibody to the
CD5 antigen on T-lymphocytes has also been reported (Kernan et al.,
1988). Isotopically labelled antibodies to tumour antigens have been
used to image tumours in vivo (Buchegger et a!., 1983) and studies are
now in progress to determine the effectiveness of isotope-labelled
10

antibodies in tumour therapy (Byers and Baldwin, 1988). Unmodified
antibodies may also be useful as therapeutic agents and some success in
the treatment of B-cell tumours using anti-idlotypic monoclonals has
been reported (Miller et al., 1982; Lowder et al., 1987). Unmodified
anti-T-cell receptor antibodies (OKT3) have also been used to prevent
renal graft rejection presumably by preventing lymphocyte-antigen
interactions (Goldstein et al., 1985).
Most clinical trials to date have used murine monoclonal antibodies
(Byers and Baldwin, 1988) but successful treatment has so far been
limited by the generation of an immune response to murine antibody
which often precludes continued treatment (Levy and Miller, 1983;
Jaffres et al., 1986). The use of human monoclonal antibodies would
alleviate this problem but is limited by the difficulties in producing
human antibodies to tumour antigens (James and Bell, 1987). Recent
development of chimeric antibodies containing either human constant
and mouse variable regions (Morrison et al., 1984; Boulianne et al., 1984;
Morrison, 1985) or rat hypervariable regions inserted into a human
antibody (Riechmann et al., 1988) may soon overcome these problems.
A further potential application for monoclonal antibodies is in their
use as anti-idiotype vaccines. Anti-idiotypic antibodies which are
raised against the antigen binding site of a second antibody, can mimic
the conformation of the antigen recognised by the second antibody. Such
antibodies thus have the potential to act as immunogens and could be
used as vaccines (Reagan et a1., 1983; Kaufmann et al., 1985).
The realisation of the potential of monoclonal antibodies for
diagnostic and therapeutic products will require the production of large
amounts of antibody. Diagnostic assays may' only require microgram
quantities of antibody for individual tests but could consume gram
11

quantities of antibody if used in large scale screening such as in ABO
blood typing (Voak and Lennox, 1983). In vivo imaging applications and
immunotherapy may require tens of kilograms assuming that each
patient may require several hundred micrograms for imaging (Mach et
al., 1981) and several hundred milligrams during treatment (Miller et al.,
1982). Furthermore as the economics for large scale production of
monoclonal antibodies improve then kilogram quantities of antibodies
could be consumed in biochemical purification processes. It has
therefore been predicted that the market for monoclonal
antibodies will increase substantially in the next decade (Ratafia, 1987;
Spier, 1988) Illustrating a real need for an efficient production system
to meet such a demand.
1.4 Production of monoclonal antibodies.
1.4.1 A scites.
Hybridomas can be grown as tumours in the peritoneal cavity of
histologically compatible mice and the antibody harvested from ascitic
fluid at concentrations of 5-20 mg/ml. A single mouse might produce up
to 15m1 of ascitic fluid before succumbing to the tumour (Hurrell, 1982).
There are however several disadvantages in the use of rodents to
expand monoclonal antibodies production.
1) Mouse ascites may contain contaminating antibodies from the host.
Other detrimental proteins such as proteases may also be present.
ii) Mouse adventitious agents may be present in ascitic fluid.
iii) Human hybridomas are difficult to propagate in rodents.
iv) Large scale applications would require large numbers of mice. The
lack of the economy of scale and process control would have to be
12

considered in addition to any ethical objections to such a production
process.
For these reasons many workers are investigating the in vitro
production of monoclonal antibodies and because hybridomas grow
readily in suspension culture many of the culture techniques used for
the propagation of microorganisms have been applied to the large scale
culture of hybridoma cells.
1.4.2 invitro production methods.
The stirred tank reactor has already been used extensively for
both research and production scale cultivation of mammalian cells. the
largest to date being an 80001 reactor for the production of
lymphoblastoid interferon (Phillips et al., 1985). As yet the use of large
scale homogeneous stirred tank systems for the production of monoclonal
antibodies is limited but Celltech have pioneered the use of a 10001
airlift reactor for the commercial production of monoclonal antibodies.
Using this system Birch et al. (1985a) claim to produce an average of
150g of antibody in a typical run. This fermenter has now been scaled up
to 20001 and is used for the production of both antibodies and
recombinant proteins by mammalian cells (Rhodes and Birch, 1988).
The concentration of product in a homogeneous suspension culture
is limited by the maximum cell density that can be achieved in such a
system before either nutrients are exhausted or toxic products become
inhibitory. This is typically 1-2 x 106 cells/ml for a hybridoma and
antibody titres ranging from, 20-200 jig/ml can be obtained (see for
example, Reuveny et al., 1985; Birch et aL. 1985a; Fazekas de St Groth,
1983). However, increased cell densities may be achieved by perfusing
13

the culture with fresh medium. The increased nutrient supply and the
removal of toxic metabolites extends the production phase of a culture.
Himmelfarb and colleagues (1969) showed that higher cell densities
could be obtained in perfused culture in which cells were retained by
removing waste medium through a spinning filter device. The high shear
force generated at the surface of the spinning filter alleviated the
blocking which occured with static filters.
Variations on the spinning filter device have been adopted for the
production of monoclonal antibodies (Tolbert et a)., 1985; Reuveny et
al., 1986b). These authors have reported hybridoma densities of up to
3x107/ml and increased antibody titres in the effluent stream (50-500
fg/ml). However filter blockage still occured after several days of
culture and it was necessary to replace the filters regularly to maintain
long term cultures spanning several weeks.
1.4.2.1 Immobilised cell systems.
The relative fragility of animal cells and the low product
densities experienced in homogeneous suspension cultures has led to
research into the use of immobilised cell systems for high density cell
culture. Immobilised cells are not exposed to the hydrodynamic stresses
experienced In agitated systems and are separated from the medium
which facilitates medium perfusion and downstream processing.
1) The entrapment of cells within porous matrices.
A number of techniques have been applied to the immobilisation
of cells in porous beads using materials such as alginate (Nilsson and
Mosbach. 1980) or agarose (Nilsson et al.. 1983; Katinger and Sheirer.
1985). These materials retain cells but nutrients and products may
14

diffuse freely.
Although hybridoma cells are non-adherent, materials with more
open pore structures have also been successfully used for hybridoma
immobilisation. In these systems it appears that cells settle within the
tortuous network of channels and subsequent outgrowth prevents the
cell mass becoming dislodged. Hybridoma densities approaching 108
cells/cm3 of matrix were observed in fibrous collagen beads (Karkare et
al, 1985). A variety of materials including polymer foams have also been
investigated for their effectiveness In entrapping hybridoma cells
(Lazar et a!.. 1987, Murdin et a!.. 1987). Packed beds were used in these
investigations but problems with mass transfer due to channeling and
cell outgrowth have been reported in packed beds (Karkare et al, 1985).
Karkare and colleagues therefore favour a fluidised bed system using
weighted collagen beads which allow higher medium flow rates and
improved mass transfer. Bubble columns, air lift and stirred reactors
have also been used for the cultivation of immobilised cells (Himmler et
al., 1985).
A different approach utilises a ceramic matrix perforated by a
number of uniform square channels (Lydersen et al., 1985). Hybridomas
proliferate in the porous walls of the ceramic matrix and medium is
supplied through the channels allowing the system to operated for
extended periods.
All the systems described above allow free passage of the product
into the medium and result in product titres which are often no higher
than those achieved in homogeneous suspension culture.
ii) Encapsulation.
Cellular microencapsulation using poly-l-lysine membranes was
first described by Lim and Sun (1980), who used encapsulated islets as
16

an artificial pancreas. Microencapsulation has since been used for a
number of cell lines (Jarvis and Grdina, 1983).
Damon Biotech Incorporated have adopted this technology for the
production of monoclonal antibodies (Rupp, 1985). The use of a
poly-l-lysine capsule with a molecular weight cut off of 8OkD means
that antibody is retained within the capsule while low molecular weight
nutrients and waste products can diffuse freely across the membrane.
The advantages of this system are that cells are protected from
environmental stress and antibody reaches high concentration (5mg/mi
capsule volume). As this may represent up to 50% of the total
Intracapsular protein, product recovery during downstream processing
is therefore improved. However there are disadvantages to such a
system, which is essentially a batch process, because the capsules must
be ruptured to harvest the antibody. The manufacture of microcapsules
is a complex process and the scale up of such a system would be limited
by the capacity to manufacture large numbers of microcapsules.
iii) Hollow fibres.
The use of capillary systems for the growth of animal cells was
first described by Knazek et al. (1972) and subsequently developed by
Amicon. In this configuration cells are grown in the spaces between a
tightly packed bundle of hollow fibres. Medium is pumped through the
porous fibres and nutrients diffuse
, into the, extracapillary space.
Likewise waste metabolites diffuse into the capillaries and are removed
by the medium stream. Cells can be grown to high density (> 108 /ml) in
conditions which are similar to those experienced by tissues in vivo.
Several cell types have been successfully cultured in these systems
(see review by Hopkinson, 1985) and since 1981 the hollow fibre sytem
has been used for both human and mouse monoclonal antibody
16

production (Calabresi et al.. 1981; Wlemann et al., 1983; Altshuler et al..
1986; Schonherr et al., 1987). High concentrations of antibody can be
achieved in the extra-capillary space when low molecular weight cutoff
(10-100 kDaltons) membranes are used and unlike microcapsules
antibody can be harvested continuously or intermittently from the
extra-capillary space. The process is therefore continuous and can be
operated for several months.
The hollow fibre systems described above have a closed shell and
medium and waste products across the membrane by diffusion. An
important limitation of this configuration is the formation of nutrient
and metabolite gradients along the length of the fibre bundle due to high
cell densities (Ku et al., 1981) and due to the pressure gradient which
exists in such systems (Tharakan and Chau, 1985a). These effects can
be reduced to some extent by increasing flow rates and, periodically
reversing the direction of flow (Hopkinson, 1986). Improved mass
transfer has been reported in cross flow systems where medium is
delivered from the shell side and exits through the fibre lumen (Ku et
aL, 1981; Tharakan and Chau, 1985b) while In other systems the_ use of
pressure cycling ensures that medium passes back and forth across the
fibre wall and hence give mass transfer rates higher than that of
diffusion alone.
Other reactors use ultrafiltration membranes as in hollow fibre
reactors but the cells are sandwiched between flat membranes (Seaver
and Gabriels, 1985). Klement et al. (1987) have modified the design
further to incorporate different porosity membranes which allow the
separation of cells, medium and product into different compartments.
Such systems could incorporate several layers of membranes to facilitate
scale-up.
17

1.4.3 Assessment of bioreactor types.
The type of bioreactor employed for a particular application
depends on the properties of the cell and the scale of the process.
Several authors have reviewed aspects of fermenter design for the scale
up of animal cell production processes (Glacken et al., 1983; Lambe and
Walker, 1987; Katinger and Scheirer, 1985; Hu and Dodge, 1985; Merten,
1988). Merten has assessed reactor design based on the properties of the
cell line namely. shear sensitivity, production kinetics and genetic
stability and these parameters will be discussed for hybridoma cells.
V Shear sensitivity.
Fazekas de St Groth (1983) found that hybridoma cell doubling
times in stirred suspension culture increased as agitator speeds
approached 100 rpm but noted some variation in the sensitivity of
different hybridoma lines to high stirrer speeds. Other authors have
found little effect on hybridoma growth at stirrer speeds in excess of
100rpm (Boraston et al., 1984; Dodge and Hu, 1987). This suggests that
many hybridoma lines are less sensitive to shear than has been supposed
and are thus amenable to suspension culture techniques.
ii) Production kinetics.
The production kinetics for a particular protein will have a
profound effect on the design of a production process depending on
whether the product is growth-associated, non growth-associated or
subject to feedback regulation. Feedback regulation would preclude the
use of product retention systems described above. Growth-associated
production kinetics would be favoured by rapid growth rates precluding
18

the use of immobilised cell systems where growth is restricted. However
in hybridomas the production of antibody is not growth associated and
antibody production continues in the absence of cell proliferation (Birch
et al., 1985b; Miller et al., 1988a). Furthermore, there is no evidence of
feedback regulation of antibody production in hybridomas (Fazekas de
St Groth. 1983). These production kinetics would be compatible with high
cell densities and product retention systems.
iii) Genetic stability.
Merten (1988) has suggested that cell retention systems which
restrict cell growth but allow extended production periods would be
preferable for genetically unstable cell lines. The loss of genetic
material during cell replication could be minimised In such systems.
The stability of hybridoma lines varies considerably but several
hybridoma lines have been maintained in culture for long periods of time
without any apparent changes in the production characteristics (see for
example Low and Harbour, 1985a; Birch et al. 1985b; Bree et al., 1988).
It would appear that the properties of hybridoma cells do not
preclude any of the production methods described and the design of the
process must ultimately
be centred upon the economics of the process.
Stirred suspension culture systems are unit processes and benefit from
the economy of scale, but low product concentrations increase the
complexity of downstream processing. The high density product
retention systems yield a more concentrated product which improves
product recovery but scale up of such processes is often limited to an
increase in the number of units and thus economy of scale cannot be
realised. However cells grown at high density may have a reduced
requirement for certain medium components such as serum (Hopkinson,
1985; Reuveny et al., 1987) and thus make a more economical use of
19

medium. The high cost of mammalian cell culture medium suggests that
any improvement of cell or product yields by process control and/or
medium design will yield more economical production systems. Such an
approach requires knowledge of antibody production kinetics and cell
physiology and some aspects of these are discussed in the following
section.
1.5 Hybridoma growth and production kinetics.
Growth kinetics for hybridomas appear to follow the characteristic
bacterial growth kinetics because there is a discernable lag phase
followed by exponential growth and decline phases although cell
doubling times are generally longer, typically 12-24 hours for mouse
hybridomas.
The lag phase may be due in part to the population density
dependence of cell growth observed in animal cell culture. This was
shown by Eagle (1962) to be a dependence upon a pool of metabolic
intermediates which cells must synthesise before they can divide. The
result is that there is a minimum population which will condition the
medium sufficiently to support cell growth. Eagle demonstrated that by
adding various metabolic intermediates the population density
dependence could be reduced.
Maximum cell densities achieved in static or stirred batch cultures
is generally 106 cells/ml (Boraston et al., 1984; Reuveny et al., 1987)
but this figure can be doubled by batch feeding and improved medium
design (Rhodes and Birch, 1988).
Various observations on the production kinetics for monoclonal
20

antibodies suggest that the antibody production rate is independent of
growth rate. For example a number of authors have observed- that
antibody production continues into the stationary phase and In a
number of cases It has been observed that specific antibody production
rates are maximal during the stationary phase (Fazekas de St Groth,
1983; Boraston et al., 1984; Reuveny et al., 1985; Birch et al., 1986). In
contrast some hybridoma lines show a decreasing specific production
rate as they approach stationary phase (Lavery et al., 1985) and are
characteristically low producers. The latter production kinetics may be
due to some inhibition mechanism as postulated by Merten et al. (1985)
who showed that the addition of spent medium to the culture greatly
reduces the specific production rate of antibody in these cells. This
effect seemed to be highly specific in that it could not be reproduced
using spent media from other cultures or using purified mouse
monoclonal antibody from other hybridomas. By contrast Fazekas de St
Groth (1983) has examined a number'of hybridoma lines in continuous
culture and has found no evidence of feed back Inhibition of antibody
synthesis because the specific production rate for antibody' was
essentially the same at cell densities between 105 and 106 /ml. '
The underlying mechanism of the former production kinetics has
posited two possible explanations, firstly that an intracellular pool of
immunoglobulin builds up in intact cells and is released during the death
phase as cell membranes are disrupted and secondly that the specific
production rate increases during the stationary phase. There is
conflicting evidence in that Emery et al. (1987) claim to have detected
an increasing intracellular pool of immunoglobulin in some cell lines
whilst Birch et a!. (1987) found no such evidence in their cell lines.
Much of the evidence would tend to favour an increased production rate
21

during the decline phase. It has been shown for example in perfused
culture systems that a stationary cell population is achieved in which
the specific antibody production rate is sustained at a higher level than
during the growth phase (Van Wezel et a!., 1985; Reuveny et a!., 1986).
It has also been shown that the antibody production rate is related to
the number of viable cells (Velez et al., 1986; Renard et a!.. 1988) with
an insignificant contribution from dying cells.
It is evident from the widely ranging antibody productivities seen
in different hybridoma cell lines, that the production characteristics for
a hybridoma is ultimately determined by the cell line itself. This results
from a combination of the parental genotypes (ie myeloma and B-cell),
the differentiation state of the B-cell and any chromosome
rearrangement or recombination events which take place after cell
fusion (Westerwoudt, 1986). Thus it is possible that each hybridoma line
possesses a unique phenotype which may be manifested as defects in cell
metabolic pathways or altered production kinetics. In addition defects
in the metabolism of theýmyeloma parent may be complemented by the
other fusion partner. Such a phenomenon has been noted for the NS-1
myeloma which has a requirement for lipids in serum-free medium
whereas hybridomas derived from the NS-1 parent do not have this
requirement (Kawamoto et al., 1986). The influence of the parental cell
phenotype on the growth and productivity of the resulting hybrid has
not been rigorously examined but it is likely that optimisation of the
medium and other conditions for the production of antibody will have to
be undertaken for each hybridoma line. However the effects of the
environment on the growth and metabolism of cultured cell lines in
general and hybridomas in particular will be discussed in the next
section in an attempt to identify those factors which Influence antibody
22

production.
1.6 The effect of the environment on cell growth and antibody
production.
1.6.1 Carbohydrates and cell growth.
A general review of carbohydrate metabolism may be found in
the paper by Morgan and Falk (1986) in which they list carbohydrates
which have been shown to support cell 'growth and discuss the major
regulatory enzymes of carbohydrate metabolism.
Rapidly dividing cells exhibit a high rate of glycolysis and a
substantial proportion of glucose is metabolised to lactate even under
aerobic conditions (Warburg. 1956; Levintow and Eagle, 1961; Roos and
Loos, 1973; Reitzer et al.. 1979) and this is also true of hybridoma cells
(Low and Harbour 1985; Hu et al.. 1987; Luan et al.. 1987). This was
initially thought to be a characteristic of tumour cells (Warburg, 1926)
and indeed there is evidence that the regulation of phosphofructokinase
(a key regulatory enzyme in glycolysis) may be altered in transformed
cells (Morgan and Falk. 1986) possibly due to a reduced sensitivity to
feedback inhibition by ATP (Eigenbrodt et al.. 1985). However certain
rapidly proliferating normal cells and cells with the potential for rapid
proliferation (e. g. lymphocytes), also exhibit high rates of glycolysis.
Likewise it has been found that there is no direct correlation between
glucose uptake and cell yield (Eagle. 1958). This phenomenon is perhaps
now resolved by recent studies concerning the contribution of other
carbon sources, in particular glutamine, to energy metabolism and this
subject is discussed further in a later section.
It has been suggested that a possible role for the apparently
23

inefficient rate of glycolysis proliferating cells is the maintenance
of elevated concentrations of glycolytic intermediates, for
macromolecular synthesis (Hume et al., 1977). Reitzer et al. (1979)
have gone further and suggested that the primary role of glycolysis In
HeLa cells was to provide precursors for nucleotide synthesis via
the pentose phosphate pathway. They subsequently showed that
several cell lines would grow in the absence of glucose provided uridine
or cytidine were present (Wice et al.. 1981).
It has been observed that high glucose concentrations increase
both the glucose consumption rate and the yield of lactate from glucose
in several cell types (Graff. et a!., 1966; Renner et al., 1972; Glacken et
al., 1986) Including hybridomas (Low and Harbour. 1985b; Hu et al.,
1987; Luan et a1., 1987). Renner et al. (1972) found that at glucose
concentrations less than 60pM, all of the glucose was used for
macromolecular synthesis and oxidative processes with no net formation
of lactate. This may be due to a change in the flux of glucose through
different metabolic pathways as proposed by Reitzer et al., (1979). High
concentrations of glucose lead to an increased flux through glycolysis
with the subsequent production of large amounts of lactate whereas at
low glucose concentrations the flux through glycolysis is greatly
reduced with a greater proportion of glucose carbon being diverted
through anabolic pathways. The increased rate of glycolysis at high
glucose concentrations is accompanied by a decreased oxygen
consumption rate (Frame and Hu, 1985; Glacken et al., 1986) possibly
due to the increased production of ATP . from , glycolysis inhibiting
oxidation in Krebs cycle (Miller et a!., 1988a). High growth rates also
lead to an increased rate of glucose consumption and increased yields
of lactate from glucose (Hu et al., 1987, Luan et a!., 1987b).
24

The production of large amounts of lactic acid can quickly
overcome the buffering capacity of cell culture media leading to a drop
in pH and inhibition of cell growth. Eagle et al. (1958) found that the
substitution of more slowly metabolised sugars such as fructose or
galactose for glucose, leads to a greatly reduced lactate yield. Reitzer
et al. (1979) subsequently demonstrated that most of the fructose
utilised by HeLa cells was metabolised via the pentose phosphate
pathway with very little accumulating as pyruvate or lactate. The
substitution of glucose by fructose or galactose has therefore been used
by a number of workers to reduce the lactate concentration in culture
media and improve pH control (Imamura et al., 1982; Reitzer et al., 1979;
Low and Harbour, 1985b).
1.6.2 The effect of glucose and lactate on antibody production.
Products of glucose catabolism have been found to repress
biosynthetic pathways in lower organisms and there is some evidence
that catabolite repression may also occur in mammalian cells. For
example higher titres of interferon were achieved in glucose-limited
chemostat culture of mouse LS cells than in batch cultures with high
glucose concentrations (Tovey eta)., 1973). The production of interferon
from Nalmalva cells was also found to be inhibited by increased glucose
concentrations in chemostat culture (Kromer and Katinger. 1982).
Studies on glucose-limited chemostat cultures of hybridomas have
revealed that the specific antibody production rate is highest at low
growth rates which coincides with the lowest metabolic quotient for
glucose (Birch et al.. 1985b; Miller et al., 1988a). Such an effect would
be consistent with catabolite repression but there is evidence to suggest
25

that this is not the case in hybridoma cells.
1. Low and Harbour (1985b) found that high glucose concentrations do
not inhibit antibody production in batch culture. These findings are
supported by Tharakan and Chau (1986c) who found no apparent
relationship between glucose uptake rate and antibody production rate.
2. Substitution of more slowly metabolised carbohydrates such as
fructose for glucose does not improve antibody titres (Low and Harbour,
1985b).
3. Miller et al. (1988a) found that at constant growth rate the highest
metabolic quotient for antibody production coincided with the highest
metabolic quotient for glucose uptake, an effect which is not consistent
with catabolite repression.
The inhibitory effect of lactic acid on cell growth is probably due
to the lowering of the culture pH to suboptimal levels although it has
been suggested that the lactate ion per se inhibits hybridoma growth
and antibody production independently of pH (Glasken et al.. 1986).
Others (Reuveny et a!., 1986a; Miller et al., 1988b) have found that
lactate concentrations in excess of 25mM exhibit little effect on either
cell growth or antibody production. The sensitivity of cells to lactate
toxicity may therefore vary considerably between cell lines.
1.6.3 Amino acid metabolism.
Early studies on the amino acid requirements of animal cells were
largely empirical being based either on the composition of biological
fluids (Fischer. 1948; White, 1949; Morgan et al. 1950) or by the
evaluation of the effect of individual amino acids on cell proliferation
(Eagle, 1955; Eagle, 1959). Later studies on changes in amino acid
26

concentrations during cell growth have shown that patterns of uptake
or accumulation of amino acids in culture media vary with the cell type,
medium composition and proliferation state. However, in general it
appears that glutamine is the most rapidly utilised followed by leucine,
isoleucine, lysine and valine (McCarty, 1962; Kruse et al., 1967.
Griffiths and Pirt, 1967; Stoner and Merchant, 1972; Roberts et al.,
1976). When non-essential amino acids are omitted from medium they
are synthesized by cells during growth and may accumulate in the
medium (McCarty. 1962; Griffiths and Pirt. 1967; Stoner and Merchant,
1972) but are consumed during the stationary phase (Stoner and
Merchant. 1972). When non-essential amino acids are included in the
medium they are utilised by the cells (Pasieka et al., 1960; Kruse et al.,
1967) and the consumption of essential amino acids is reduced (Griffiths
and Pirt, 1967). There are few reports concerning the amino acid
requirements of hybridoma cells but it has been reported that there is
rapid consumption of Blutamine, arginine and serine with isoleucine,
leucine, methionine, valine. phenyalanine and tyrosine being used to a
lesser extent (Merten et al.. 1986). The rapid consumption of serine may
reflect a requirement for serine in common with certain other cell lines
(Kruse eta).. 1967; Stoner and Merchant, 1972; Birch and Hopkins, 1977),
serine being an important intermediate in the synthesis of purines and
pyrimidines via tetrahydrofolate and glycine.
The importance of glutamine for the growth of animal cells in
culture has long been known (Eagle et al., 1956, Kitos et al.. 1961) and
its role in cell metabolism has been reviewed recently by McKeehan
(1985). The rapid utilisation of glutamine appears to be a characteristic
of proliferating cells such as tumour cells (Kvamme and Svenneby,
1961; Kovacevic and Morris, 1972; Regan et a)., 1973) or rapidly
27

dividing normal cells . Glutamine is also consumed rapidly by mouse
myeloma cells (Roberts et al., 1976), lymphocytes (Ardawi and
Newsholme, 1982) and hybridomas (Seaver et al., 1984). This
requirement for glutamine is due In part to the central role that
glutamine plays in many biosynthetic pathways. Glutamine is the
amino donor in the synthesis of purines and pyrimidines, amino
sugars, pyridine nucleotides and asparagine in mammalian cells. Indeed
it has been shown that the availability of glutamine regulates the
rate of synthesis of purines and pyrimidines and hence cell growth
(Fontanelle and Henderson, 1969; Ley and Tobey, 1970; Zetterberg and
Engstrom, 1981).
Glutamine carbon has been shown to accumulate as glutamate and
aspartate with lesser amounts of alanine, pyruvate and lactate in
lymphocytes (Ardawi and Newsholme, 1982a). Reitzer et al. (1979),
discovered that lactate was a significant product of glutamine
metabolism accounting for up to 13% of glutamine carbon in HeLa cells.
Hybridoma cells produce large amounts of alanine during cell growth
(Seaver et aL, 1984; Merten. 1986) suggesting that alanine transaminase
activity is high in these cells and may account for a significant
proportion of glutamate nitrogen. Figure 1.2 shows the possible fate
of glutamine in hybridomas.
There is also considerable evidence to suggest that glutamine
is an important energy source for animal cells and the appearance
of large amounts of glutamine carbon as C02 indicates a high rate of
glutamine oxidation. Lavietes et al. (1974) showed that glutamine
oxidation accounted for 70 to 80% of oxygen uptake in lymphoma cells
28

m
ö
v
ZOQOZ
LA
Ny.
ý
ösm0
WZý_
W
LLJ
cr: l (a C, fc-11
m
d
R- cd
.a .
.. r
G
0
G
Co 10
f4.4
N ý
CLI
m
Co 0
94
Sri
0 90 i G
4 gz d Co c3 i
Ö
4.1
ö
, ýv ý
d
CO
a. .,
Co w
C) b
" 4a
rw r.
4)
ce
Ici
ca C: )
r= CD X
CD
:i ro 4..
CD
N CC E E E E
Co
4.3 yea
ä ;j ý
a. . Co
,m
r+ N CJ 'd'
,
d
Ci
Co
a
(D
r.
W
0
Co
Z.
CI)
4. -1 4.1
n
. 41
10 CO - L-
29
td
q
0.4 V
cis Co
cät gal
m
Co
a
o o
Co
Y
a)
to
e.
>.
L
a ts4
" CO Q) - -4
N
CL
O
,r
E) :5
O
ýij
5-1
03

and Donnelly and Scheffler (1976) found that the oxidation of
glutamine could provide 40% of the energy requirements in Chinese
hamster ovary cells. The work of Reitzer et al. (1979), on HeLa cells
suggested that in the presence of glucose, glutamine oxidation via the
TCA cycle provided greater than 50% of the cells energy requirements
whilst 80% of glucose was converted to lactate and only 5% of glucose
carbon entered the TCA cycle. Furthermore when glucose was replaced
by fructose almost all the fructose carbon passed through the pentose
phosphate cycle and they suggested that glutamine oxidation provided
98% of the cells energy requirements. However Lazo (1981) has
suggested that whilst glutamine may be the preferred respiratory
substrate in many cells, aerobic glycolysis may still provide much of the
cells energy requirements in rapidly growing cells.
A number of authors have shown an apparent reciprocal
regulation of glycolysis and glutamine oxidation. Kvamme and
Svenneby (1961) noted that glutamine oxidation was inhibited by
glucose in Erhlich ascites tumour cells and Zielke et al. (1978) found
that glutamine utilisation in human diploid fibroblasts was regulated
by glucose concentration and vice versa. A similar phenomenon has
been noted in hybridomas by Hu et al. (1987) who found an increase in
glutamine oxidation in low glucose media. Miller et al.. (1989a) have also
observed that pulse addition of glucose to steady state hybridoma
cultures leads to a decrease in the glutamine consumption rate and a
reduced rate of oxygen uptake. They suggested that the increased ATP
to ADP ratio due to a higher rate of glycolysis was responsible for
inhibition of oxidation in Krebs cycle. Alternatively, Glacken (1988) has
postulated that the increased ATP production from glycolysis lowers the
intracellular phosphate pool which in turn lowers phosphate-regulated
30

glutaminase activity assuming that glutaminase is of the
phosphate-dependent type as has been suggested for many cultured
cells (McKeehan, 1986). The mechanisms of this reciprocal regulation
have yet to be elucidated and the subject is further complicated by
Ardawl and Newsholme (1982) who showed that glutamine uptake is
stimulated by glucose in proliferating lymphocytes. However it is
likely that the changes in these major metabolic pathways will have
profound effects on energy metabolism and biosynthesis and are thus
important areas for study when considering the optimisation of antibody
production in hybridoma cells.
1.6.4 The effect of Blutamine on antibody production.
Crawford and Cohen (1985) found that not only was glutamine
essential for lymphocyte differentiation in vitro but also that
immunoglobulin synthesis and secretion was considerably enhanced in
the presence of glutamine.
The effect of glutamine concentration on antibody
production by hybridomas has not been extensively studied. Birch et a).
(1986b) have used a glutamine-limited chemostat culture to study the
effect of cell growth rate on antibody production. They found that the
antibody production rate increased slightly with increasing growth
rate. This may be a result of the increased uptake and presumably
higher intracellular concentrations of glutamine at high growth rates.
This may stimulate antibody production which would be consistent with
Crawford and Cohens observations perhaps by providing a larger amino
acid pool for protein synthesis. However metabolic data was not
discussed in Birch et als paper and it is not possible to determine the
31

effects of glutamine on cell metabolism from their data. Hence the
relationship between glutamine metabolism and antibody production
remains obscure. Other studies have been concerned with prolonging
cell viability and reducing ammonia accumulation by feeding glutamine
(Luan et al.. 1987a, 1987b; Glacken et al., 1986; Reuveny et al., 1986b)
and although this has led to improved cell yields and hence higher
antibody titres, the effect of glutamine on the specific production rate
of antibody has not been investigated.
It has not yet been determined if the changes in cell metabolic
pathways associated with different energy sources or cell growth rate
affect protein synthesis or secretion in hybridomas. Such information
would facilitate the design of improved media for immunoglobulin
production and the design of production processes.
1.6.5 The effect of ammonia on cell growth and antibody production.
Ammonia is a major product of glutamine metabolism due to
glutaminase activity. Griffiths & Pirt (1967) found that the
incorporation of nitrogen into biomass was very inefficient when
glutamine was present but was almost 100% efficient when cells were
adapted to grow on glutamate suggesting that glutamine is the main
source of ammonia in culture media. The rate of glutamate oxidation As
low (Lazo, 1981) and - it has been suggested that glutamate
aminotransferases are more important than glutamate dehydrogenase in
the conversion of glutamate to a-ketoglutarate (Glazer et a!.. 1974;
Ardawi and Newsholme. 1982; Moreadith and Lehninger, 1984) resulting
in a more efficient incorporation of nitrogen and lower ammonia levels.
There is also some evidence that exogenous glutamate is metabolised
32

differently to glutamate derived from glutamine. It appears that
exogenous glutamate is more likely to be converted to a-ketoglutarate
by transamination reactions whereas glutamate derived from glutamine
Is oxidatively deaminated by glutamate dehydrogenase (Borst, 1962;
Kovacevic, 1971; Schoolwerth and LaNoue, 1980). This may be an effect
of compartmentation because both glutaminase and glutamate
dehydrogenase are located within the mitochondrion whereas the
transaminases are located predominantly in the cytosol.
The rate of ammonia production is also increased at high
glutamine concentrations and the controlled feeding of glutamine to
maintain low levels of ammonia was successfully employed by Glacken
et al. (1986).
Ammonia has been shown to inhibit cell growth at levels above
2mM in several cell lines (Visek et al., 1972; Holley et al., 1978; Butler,
1985) including hybridoma cells (Reuveny et al., 1986a; Dodge et al.,
1987). Ammonia has also been shown to inhibit the production of viruses
(Eaton and Scala, 1961; Jensen and Liu, 1961) and interferon (Ito and
McLimans, 1981) and its inhibitory effect on hybridoma growth is
reflected in lower antibody titres achieved in media containing high
levels of ammonia (Reuveny et al., 1986a). Whether the effect Is due to
inhibition of immunoglobulin synthesis per se or a reflection of a general,
depression of cell metabolism caused by ammonia has not been
ascertained but the latter seems more likely.
1.6.6 Oxygen requirements by cells in culture.
Animal cells in culture have been reported to have oxygen
uptake rates of between 0.05 and 0.5 pmoles/106 cells/hour (Fleischaker
33

and Sinskey. 1981). The oxygen uptake rate for hybridoma cells falls
within this range with a maximum uptake rate of 0.2 pmoles/106
cells/hour (Boraston et a1.. 1984; Miller eta!.. 1987). The oxygen uptake
rate also appears to be related to cell growth rate (Katinger. 1987) and
has been used as an indicator of the metabolic state of cells in
immobilised cell systems (Lydersen, 1985; Pugh, 1988). However the
observation that oxygen uptake is inhibited by high glucose
concentrations (Crabtree, 1929; Frame and Hu, 1984; Glacken et al..
1986; Miller et al.. 1988) means that care must be taken when
correlating oxygen uptake with cell growth rate.
The optimum oxygen concentration for cell growth varies with cell
type but is generally about 60% air saturation (Kilburn and Webb, 1968;
Radlett et al., 1972). A similar optimal concentration of oxygen for
hybridomas has been noted by some authors (Reuveny et al., 1986a;
Phillips et al., 1987) whilst Boraston et al. (1984) found little effect on
growth rate for oxygen concentrations between 8 and 100% air
saturation. Oxygen concentrations greater than those of atmospheric air
have been reported to be toxic to mammalian cells (Cooper et al., 1958;
Brosemer and Rutter, 1961; Mizrahi et al., 1972) although the
mechanisms for such toxicity have not been elucidated.
Cell damage by air bubbles has also been reported by a number of
authors (Telling and Esworth, 1965; Kilburn and Webb, 1968; Radlett et
al.. 1972; Spier, 1982) which Handa and colleagues (1987) suggest is due
to the energy dissipation in the region of bubble disengagement. For this
reason various devices have been described which provide bubble free
aeration by the use of silicone tubing (Kuhlmann, 1987). porous
membranes (Lehmann et al.. 1987) or caged aerators (Whiteside and
Spier. 1985). High concentrations of serum and surface active polymers
34

have been reported to protect cells from bubble damage (Kilburn and
Webb, 1968; Handa et al., 1987), however Birch and colleagues have
successfully cultured several cell types (including hybridomas, Rhodes
and Birch, 1988) in airlift reactors using low protein media and have not
reported any such cell damage due to bubbles.
1.6.7 The effect of oxygen on antibody production.
Mlzrahi and colleagues (1972) studied the effects of oxygen on
antibody production by a human lymphocyte cell line and found that
specific antibody, production rates were highest at oxygen
concentrations which were below the optimum for cell growth.
Some hybridomas also show a similar phenomenon in that optimal
growth occurs at 60-60% air saturation whilst optimal antibody
production occurs at 15-25% air saturation (Reuveny et a1., - 1986a;
Phillips et al., 1987) and Birch et al. (1985) found that antibody
production rate was higher at low growth rates in oxygen limited
chemostat culture. These observations suggest that high oxygen
concentrations inhibit antibody productiornllowever Miller et al. (1987)
have studied the effect of oxygen concentration on -antibody
production by hybridomas at constant growth rate and have
Indicated that optimal antibody production occurs at 50% air
saturation which coincides with the highest metabolic quotient for
oxygen uptake. In addition Bodeker et al. (1987) found that IgM
production by Immobilised hybridomas was favoured at high oxygen
consumption rates. It thus appears that the effect of oxygen on antibody
production Is not an effect of direct inhibition but rather a reflection of
the balance of cell metabolism at different growth rates.
35

1.6.8 The effect of PH on cell growth and antibody production.
The effect of pH in animal cell cultures has been measured largely
in terms of cell growth. Eagle (1973) showed' that in general, normal
fibroblasts of human or rodent origin favoured a more alkaline
environment (pH 7.4-7.9) than transformed cells from the same species
(pH 6.6-7.4) although rodent cells tolerated wider pH ranges. The
optimum pH range for hybridoma cell growth falls within the lower pH
range typically between 6.9 and 7.4 (Harbour et al., 1988).
There have been few reports on the effects of pH on hybridoma cell
metabolism although Miller et al. (1988a) have observed that the glucose
consumption rate was lower at pH 6.8 than pH 7.2 whilst oxygen
consumption remained unaltered at different pH values in chemostat
culture. Pugh (1988) also found that the rate of oxygen consumption by
hybridomas was unaffected by pH values between 6.6 and 7.2. Reduced
glucose consumption at low pH has been noted by other authors working
with different cell lines (Barton, 1971; Birch and Edwards, 1980). These
authors also found that there is a reduced rate of conversion of glucose
to lactate at low pH. a phenomenon which has also been described for
hybridomas by Luan et al. (1987b). They found that production of
lactate by hybridomas ceased when the pH had reached 6.8 but by
adjusting the culture pH to 7.2 lactate production was restored.
There are also few reports concerned with the effect of pH on antibody
production by hybridoma cells. However Miller et al. (1988a) have found
that specific antibody production rates were higher at pH 6.8 than at pH
7.2. Antibody production rates were also slightly higher at pH 7.7. The
underlying mechanisms for this phenomenon have not been explained but
36

Miller and colleagues postulated that antibody production may be
enhanced in stressed cells. However it is also likely that the changes in
cell growth rate and metabolic activity observed at low pH may favour
antibody production. The effects of pH on cell growth and antibody
production require further examination. pH is a readily controllable
parameter in bioreactors and any effect on antibody production or cell
growth has important consequences in process optimisation.
1.6.9 The role of serum in cell growth.
The use of animal sera to supplement culture media is common
practice and serum is included typically at concentrations of 6-10%.
Serum components fulfil a number of defined functions in culture (see for
example Griffiths, 1986; Maurer, 1986) as shown in table 1.2 and it is
likely that there are many components in serum which exert stimulatory
or antagonistic effects in cell culture which are as yet undefined.
However there is a limited number of serum components which are
necessary in order to achieve cell proliferation and maintain cell
viability
in the absence of serum. The use of serum free media for the
cultivation of mammalian cells is the subject of several reviews (Ham
and McKeehan, 1979; Barnes and Sato, 1980; Waymouth. 1981) and
recently a review devoted to serum free culture of hybridoma cells has
been published (Glassy et al.. 1988). The serum components which are
considered important for the growth of hybridoma cells in serum free
medium are described below.
The protein, transferrin is considered to be essential for the
growth of hybridomas (Chang et al., 1980; Murakaml et al., 1982; Kovar
37

Table 1.2. Defined components and properties of serum.
Hormaies e. g. insulin
Growth factors e. g. epidermal growth factor
fibroblast growth factor
insulin-like growth factors
Attachment factors e. g. fibronectin, fetuin
Transport/carrier proteins e. g. transferrin, albumin
Nutrients e. g. trace elements, lipids
keto-acids, amino acids,
polyamines
Buffering of pH and redcx e. g. albumin, g1utathione
Protease inhibitors e. g. 2-macroglobulin
Mechanical protection e. g. albumin, other proteins
38

and Franek, 1985). The growth stimulatory effect of transferrin appears
to be due mainly to the transport of iron to cells via specific receptors
(Aisen and Listowsky, 1980; Trowbridge and Omary, 1981) in that some
of the effects of transferrin can be reproduced with iron salts or iron
chelators (Mather and Sato, 1979; Yabe et al., 1987).
Insulin is also commonly included in serum free media and has a
number of roles. Insulin stimulates uridine and glucose uptake and the
synthesis of RNA, protein and lipid (Schubert, 1979) and increased
incorporation of glucose into fatty acids and glycogen in the presence
of insulin has been noted (Mather and Sato, 1979; Mather et al., 1981).
Optimum insulin concentrations of 5-20ug/ml have been reported for
hybridoma culture (Murakami et al., 1982; Darfler and Insel, 1982;
Kovar, 1986).
Ethanolamine was also shown to be important for the growth of
hybridoma cells in serum free medium as reported by Murakami et al.
(1982) and supported by the findings of Kovar (1986) and Tharakan and
Chau (1986). Ethanolamine is thought to be involved in membrane lipid
synthesis and the presence of ethanolamine has been shown to increase
phosphatidyl ethanolamine levels in cell membranes (Murakami et al.,
1982). Ethanolamine has also been described as having insulin-like
activity (Barseghian, 1984), increasing the rate of glucose oxidation and
lipogenesis although this may be a non-specific membrane effect.
Albumin is a common component of serum free medium and its
main function appears to be as a carrier molecule for lipids, although it
also serves as a carrier for hormones, growth factors and some trace
minerals (Barnes and Sato, 1980). Albumin may also have a role in the
detoxification of peroxide (Darfier and Insel, 1982) or excess trace
elements (Guilbert and Iscove, 1976; Iscove and Melchers, 1978). Most
39

cells can carry out the de novo synthesis of both cholesterol and many
fatty acids from acetyl CoA (King and Spector, 1981) but will utilise
lipids from culture media in which case the de novo synthesis of fatty
acids and cholesterol is suppressed (Jacobs and Majerus, 1973; McGee
and Spector. 1974; Alberts et al., 1974).
Selenium is a trace element which is frequently included in
serum-free media at concentrations of 10-100nm due to its growth
promoting ability (McKeehan et al., 1976; Guilbert and Iscove, 1976;
Murakami et al., 1981; Kovar, 1988). It appears that selenium functions
as a cofactor for glutathione peroxidase which protects cells from
damage by peroxides (McKeehan et al.. 1976).
Serum also provides a source of growth factors which have widely
differing effects on the proliferation of different cell lines (see James
and Bradshaw, 1984 for review). The presence in serum of trace minerals,
hormones and a complex pool of metabolites may also enhance cell
growth.
1.6.10 The effect of serum on cell growth and antibody production
The serum requirement by cells is measured typically in terms of
the extent of cell proliferation or relative cell growth rate. However it
has been noted that the serum requirement of non-proliferating high
density cultures such as hollow fibres and perfusion systems is
considerably reduced (Hopkinson, 1985; Reuveny et al., 1985) which may
indicate that the serum requirements for cell maintenance are less than
the requirements for cell proliferation. However it has also been
suggested that high density cell cultures create a microenvironment in
which the need for serum factors is reduced possibly due to synthesis of
40

such factors by the cells themselves (Merten, 1987).
Low serum or serum free media have been used for the growth of
many hybridoma cell lines (see review by Glassy et al., 1988). Sustained
growth and antibody production by several hybridoma and myeloma cell
lines has also been reported using a protein free medium (Cleveland et
a!., 1983). Many authors have noted an increased specific production
rate of antibody in serum deficient media in both murine hybridomas
(Baker et al., 1985; Tharakan and Chau, 1986; Tharakan et al., 1986)
and human hybridomas (Cole et a!., 1985; Glassy et al., 1987; Glassy,
1987). A similar phenomenon has also been observed in other lymphoid
cell cultures (Tanno et al., 1982; Sharath et al., 1984; Glassy et al.,
1987). This has been interpreted by some to suggest that serum contains
factors which inhibit antibody production (Tharakan and Chau, 1986,
Glassy et al., 1987) although no evidence of such an inhibition
mechanism has yet been presented. However in a recent review Glass),
et al. (1988) have shown that in general, hybridoma and lymphoid
cultures have lower growth rates in serum free media than in serum
supplemented media which suggests that immunoglobulin synthesis is
enhanced in more slowly growing cells. Further work is required in order
to clarify the relationship between serum and antibody production in
hybridomas.
1.7 Cell cycle-related changes in protein expression.
In the previous section the effects of physiological parameters on
the production monoclonal antibodies have been discussed and it has
become apparent that many of the effects may due to the regulation of
cell growth. This suggests that antibody production may be regulated
41

within the cell division cycle and there is evidence of this in other
immunoglobulin producing cell lines as will be discussed below.
The basic model for the cell division cycle is that of Howard and
Pelc (1953) which divides the cell cycle into four phases characterised
by the observable processes of DNA synthesis (S phase) and cell mitosis
(M phase) and the gaps between those processes termed Gi (the gap
between M and S) and Gz (the gap between S and M). It has been observed
that under adverse environmental conditions such as nutrient
starvation, cells tend to arrest in the Gi phase and will not resume cell
division unless favourable conditions are restored (Ley and Tobey, 1970;
Brooks, 1975; Melvin et al., 1979; Wynford-Thomas et al., 1985). This
phenomenon has also been demonstrated in hybridoma cells by the use
of flow cytometry which shows that the cells accumulate in the Gi phase
of the cell cycle as they enter the stationary phase (Schliermann et al.,
1987). Pardee (1974) proposed that there is a single point within Gi
(restriction point) at which cells are more sensitive to arrest but become
comitted to divide once beyond that point (see also review by Pardee et
al., 1986). It is now generally accepted that much of the variability in
the length of cell division cycles is attributable to variations In the
length of the Gi phase (Pardee et al., 1978) and that the duration of S,
G2 and M phases are relatively constant. It would therefore follow that
cell population distributions would vary at different growth rates and
Indeed population distribution studies of yeast in chemostat culture
have shown that the proportion of cells in the Gi phase decreases with
increasing growth rate (Scheper et al., 1987).
There are a number of proteins whose expression appears to be
related particular phases of the cell cycle (Milcarek and Zahn, 1978;
42

Bravo and Celis, 1980; Westwood et al.. 1984) and the expression of many
of these proteins can be related to specific events within the cell
division process. For example many of the enzymes involved in DNA
synthesis such as thymidine kinase, dihydrofolate reductase and
thymidylate synthase are expressed during S phase (Littlefield, 1966;
Johnson eta!., 1978; Marsani eta!., 1981; Conrad, 1971; Denhardt eta!.,
1986). There is also evidence that the regulation of histone synthesis
is closely coupled with DNA synthesis (Delisle eta!., 1983; Artishevsky
et a!., 1984; Schumperli, 1986). Expression of several of the oncogenes
is also related to specific cell cycle phases and these proteins probably
have roles in cell growth regulation (Denhardt et al., 1986).
The expression of other cell cycle related proteins may not be so
readily linked with cell division processes. For example plasminogen
activator secretion appears to be associated with the Gz phase in a rat
adenosarcoma cell line (Scott et al., 1987) and albumin production by rat
hepatoma cells was greatest during the Gl' phase (Chen and Redman,
1977). Studies on synchronised populations of various lymphoblastoid
cell lines have indicated that immunoglobulin synthesis may be
restricted to the late Gi /early S phase (Buell and Fahey, 1969,
Takahashi et al., 1969; Lerner and Hodge, 1970; Byars and Kidson, 1970;
Garatun-Tjeldsto et al., 1976; Turner et al., 1985). Other investigators
have found either no evidence of cell cycle dependent immunoglobulin
expression (Cowen and Milstein, 1972; Liberti and Baglioni, 1972;
Damlani et al.. 1979) or variable levels of expression (Killander et al..
1977). It is still not yet clear whether these effects are a result of the
non-physiological conditions used to synchronise the cells or genuine
cell-cycle dependent variation in Immunoglobulin expression. However
43

this poses a number of questions in the study of immunoglobulin
production by hybridoma cells.
1) Is immunoglobulin production restricted to the Gi phase in
hybridoma cells?
2) Could the increased rate of immunoglobulin production seen under
growth inhibitory conditions be due to accumulation of cells in the G1
phase?
3) Could variation in specific immunoglobulin production rate seen in
chemostat cultures (Birch et al., 1985; Miller et al., 1988) be
explained by variations in the duration of the Gi phase at different
growth rates?
1.8 Summary and objectives.
The use of animal cells for the production of proteins is receiving
increasing attention and monoclonal antibodies represent one such
product with considerable potential in diagnosis, therapy and protein
purification. The realisation of this potential requires that monoclonal
antibodies can be produced economically and on a large scale and the
identification of those parameters which influence antibody production
by hybridomas is a prerequisite for the optimisation of the antibody
production process.
It has become clear that several factors have apparent effects on
the rate of antibody production in hybridomas but that the underlying
mechanisms for these effects have not been elucidated. It has been
observed that some conditions which are suboptimal for cell
proliferation enhance the specific antibody production rate and studies
44

in chemostat culture have shown that antibody production rates are
often increased at low growth rates. These observations suggest that
there is an underlying relationship between cell growth rate and
antibody production. Whether this is a consequence of the changing
demands for macromolecular synthesis in rapidly dividing cells, cell
cycle regulated expression of immunoglobulin as observed in other
antibody producing cell lines or due to some repression/inhibition type
control is yet to be determined. The understanding of this relationship
will have important consequences for process design.
The aims of this study are therefore to examine the effects of
environmental factors on hybridoma growth and antibody production. In
particular the relationship between cell growth, metabolism and
immunoglobulin production will be addressed In an attempt to determine
those parameters which can be optimised for an antibody production
process.
The study of antibody production kinetics in cultured cells would be
greatly facilitated by direct on-line monitoring of antibody
concentrations. Therefore the preliminary obhective of this study was
to design an automated assay system for the real time determination of
antibody levels in hybridoma cultures.
45

Chanter 2
Materials and Methods
The materials and methods described in this chapter are those
which are applicable to all of the subsequent chapters. More specific
experimental procedures will be described in the relevant results
chapter.
2.1 Cell culture.
2.1.1 Cell line.
The hybridoma cell line used throughout this investigation was an
NS1 derived mouse-mouse hybridoma, designated 321 which secreted
IgG to paraquat (Wright et al., 1987). The cell line was a gift from Dr A.
Wright, Central Toxicology Laboratory, ICI.
2.1.2 Media
The hybridoma was routinely cultured in RPMI 1640 (Flow Labs Ltd. )
containing 2mM glutamine, 0.2% sodium hydrogen carbonate and 10%
newborn calf serum (Seralab). Filter sterilised glutamine stock solution
(200mM) was stored in working aliquots at -200 C. Sodium hydrogen
carbonate stock solution (8.8%) was autoclaved and stored at 4° C.
Newborn calf serum was stored in working aliquots at -20° C.
Antibiotics were not added to culture media.
46

2.1.3 Cell culture.
Hybridoma, cultures were seeded with 1 or 2x10' cells/ml. Cultures
seeded with 1x105 cells/ml were passaged after 3 days whilst cultures
seeded with 2x10° cells/ml were passaged after 2 days. During the
passaging procedure, cells were harvested by centrifuging at 100xg for
b minutes and resuspended in fresh medium at 370 C. Cell number was
determined by diluting aliquots of cells in 0.5% trypan blue in saline
-and counting in a modified Fuchs-Rosenthal counting chamber. Cells
which incorporated trypan blue were considered to be non-viable.
2.1.4 Cryopreservation.
Cells in early to mid exponential phase i. e. 24 hours after passage,
were harvested by centrifugation then resuspended at a density of
lx106 cells/ml in a freezing mixture of 80% RPMI 1640,10% newborn calf
serum and 10% dimethylsulphoxide (DMSO). Nunc Cryotubes were filled
with 1.8m1 aliquots of the cell suspension. The cells were then frozen
at an approximate rate of 1-3° C per minute for 1 hour then
transferred to the vapour phase of a liquid nitrogen container. The
initial rate of freezing was controlled by placing ampoules in a
polystyrene block with walls of approximately 2cm thickness in a
freezer at -700.
To check that the cells had been successfully preserved an aliquot
of cells was removed to check that they could be revived. Cells were
revived by thawing rapidly in a 370 C water bath. Once thawed the cells
Were immediately added to 10ml of prewarmed medium and then
47

centrifuged at 100xg for 5 minutes. The cells were then washed in a
further l Oml of medium to ensure complete removal of DMSO. The cells
were finally suspended in 10ml of medium in a 25cm2 plastic tissue
culture flask and incubated in a 370 C incubator with an atmosphere
containing 5% COz.
2.2 Assay techniques.
2.2.1 Glucose determination.
Glucose concentrations were determined by the o-toluidine test
based on the procedures of Hyvarinen and Nikila (1962) and Feter
(1965). and as described in Sigma test bulletin no. 635.
Samples for glucose assay were centrifuged at 100xg for 5 minutes to
remove cells and debris and stored at -200 C until required. For the
assay, 0.1ml of medium was added to 5m1 o-toluidine reagent in a glass
test tube which was then placed in a boiling water bath for 10 minutes.
A blank containing 0.1ml water and a standard containing 0.1ml of a
1mg/ml standard glucose solution were also included. The tubes were
cooled rapidly to room temperature in cold water before measuring
the absorbance of the resulting complex at 635nm against the blank.
The standard glucose sample typically gave an absorbance reading
of 0.260 ±0.010.
The assay response was linear for glucose concentrations between
0 and 2.5 mg/ml and therefore the glucose concentration in mg/ml could
be determined by dividing the absorbance of the sample by the
absorbance of the standard and multiplying by the glucose concentration
of the standard.
48

2.2.2 Lactate determination.
Lactate concentrations were determined enzymatically by
monitoring reduction of NAD during the conversion of lactate to
pyruvate by lactate dehydrogenase.
a) Reagents.
1) NAD Grade III from yeast (Sigma) 30mg/ml in distilled water.
il) Glycine-hydrazine buffer pH 9.0.11.4g glycine + 25m1 hydrazine
hydrate made up to 300m1 with water.
iii) Lactate dehydrogenase (LDH) Type X from bovine muscle 1000U/ml
(Sigma).
iv) Perchloric acid 8% w/v.
v) L-lactate standard 0.4mg/ml (Sigma).
b) Method.
Samples for lactate determination were centrifuged at 100xg for
5 minutes to remove cells and debris then deproteinised by the addition
of lml ice cold perchloric acid to 0.5m1 sample. The samples were kept
on ice for 20 minutes then centrifuged at 11,600xg for 10 minutes to
remove the precipitated protein. The deproteinised samples were
stored at 40 C until required.
For the assay 200p1 of sample was added to 2.5m1 glycine buffer.
A blank containing 200p1 perchioric acid and a standard containing
200p1 of a 0.4mg/ml (4.44 mmoles/1) lactate standard diluted 1/3 In
49

perchloric acid. were also included. NAD (200p1) was then added to
each tube and the initial absorbance at 340nm of each sample,
including the blank and standard, was measured against
glycine-hydrazine
buffer. To lml of each sample was then added 25p1
LDH and all samples were incubated at 370C for 30 minutes. The
final absorbance at 340nm for each sample was then measured. The
lactate concentration was proportional to the increase in absorbance
due to the reduction of NAD during the conversion of lactate to
pyruvate and was calculated as follows:
Lactate mmol/1 = (dS-dB) x 7.11
where dS = difference between initial and final absorbance of the
sample and dB = difference between initial and final absorbance of the
blank. 7.11 is a conversion factor based on an absorbance change of
6.22/mmole NADH and a dilution factor of 44.2.
The assay results were considered valid if the value for the
standard was within 5% of the expected value.
2.2.3 Ammonia determination.
The ammonia content of the culture medium was determined using
an EIL ammonia probe model 8002-8 (Kent Instruments). For the assay
O. lml 1M NaOH was added to 0.9m1 sample and the free ammonia thus
liberated was detected with the ammonia probe. A calibration curve
for the probe was constructed using aqueous ammonium chloride
standards in the range of 10-1 to 10-5 M and samples were read against
50

this curve.
2.2.4 Glutamine determination.
The glutamine concentration was determined enzymatically by
monitoring the liberation of ammonia during the deatnination of
glutamine by glutaminase.
a) Reagents.
1) 0.1 M sodium acetate buffer pH 4.9
11) Glutaminase grade V from Escherichia cola (Sigma) 5U/ml in sodium
acetate buffer
b) Method.
For the assay 450p1 sodium acetate buffer and 50p1
glutaminase were added to 500p1 sample. Controls were also set up for
each sample containing 500p1 sample and 500pl sodium acetate
buffer only. Samples and controls were incubated at 370C for 30
minutes. The ammonia content of each was measured as described in
section 2.2.3. The ammonia liberated due to the action of glutaminase
on glutamine was evaluated by subtracting the ammonia present in
the controls from the ammonia present in the glutaminase-treated
samples. Assuming 1 mole of ammonia was liberated from the complete
breakdown of 1 mole of glutamine the glutamine concentration of each
sample could therefore be quantified.
51

2.2.5 Antibody determination.
Antibody concentration in the medium was determined by
sandwich enzyme-linked Immunosorbent assay (ELISA). Samples for
ELISA were centrifuged at 100xg for 5 minutes to remove cells and
debris then stored at -201 C until required.
a) Reagents.
1) Wash buffer. Phosphate buffered saline (PBS), pH 7.4 containing 0.1%
Tween 20.
ii) Diluent. PBS + 10% newborn calf serum
ill) Coating buffer. 0.1M sodium carbonate buffer pH 9.6.
iv) Coating antibody. Sheep anti-mouse IgG (Scottish Antibody
Production Unit) diluted 1/1000 in coating buffer.
v) Enzyme-conjugated antibody. Sheep anti-mouse IgG-alkaline
phosphatase (Seralab) diluted 1/1000 in PBS containing 10%
newborn calf serum.
vi) Substrate for alkaline phosphatase. p-Nitrophenyl phosphate
(Sigma) 1mg/ml in coating buffer.
vii) Standard antibody. Affinity purified 321 monoclonal antibody
(see section 2.3).
b) Method.
1. Nunc Immunoplates were coated with IOOpI of coating antibody in
coating buffer and left overnight at 411 C.
52

2. The plates were washed 4 times with wash buffer.
3. Appropriate dilutions of the samples in diluent were applied to
the plate at 100jil per well and serial twofold dilutions were
carried out. The 321 standard antibody was also applied in twofold
dilutions starting at a dilution of 1/1000 (ipg/ml).
4. The plates were incubated at room temperature for 2 hours then
washed 4 times with wash buffer.
6. A 1/1000 dilution of antibody enzyme conjugate in diluent was
applied to each well at 100p! per well and incubated for a
further 2 hours at room temperature.
6. The plates were washed 4 times with wash buffer.
7. Enzyme substrate at 1mg/ml In substrate buffer was applied to all
wells at 100p1 per well. The plates were incubated for 1-1.5 hours
before reading absorbance at 405nm in a Dynatech MR600 plate
reader.
B. Titration curves of the 321 standard and samples were constructed..
The antibody concentrations of the samples were calculated from
the titration curves by dividing the titre of the sample at 50%
maximum absorbance by the titre of the standard at 50% maximum
absorbance.
53

2 .3 Purification and characterisation of the 321 antibody standard.
2.3.1 Preparation of Paraguat-AH Sepharose affinity column.
Paraquat propionate (90mg) was dissolved in 2m1 distilled water
and the pH was adjusted to 4.5.50mg 1-cyclohexyl-3
-(2-morpholinoethyl)-carbodiimidemetho-p-toluenesulphonate(CMC)
Was dissolved in lml distilled water and the pH was adjusted to 4.5. The
paraquat ligand and CMC were added to 16m1 washed and reswollen
AH-Sepharose (Pharmacia) and mixed by rolling for 18 hours. Another
26mg CMC was then added and the suspension was mixed for a further
64 hours before washing with PBS. The paraquat content of the
washings was assayed by adding 0.5ml to 2.5m1 of fresh 1% sodium
dithionite in 1M NaOH. The absorbance of the coloured product was
measured at 600nm. A standard curve for paraquat propionate was
constructed with standards ranging from 15 to 500pg/ml paraquat
propionate (figure 2.1). The amount of paraquat ligand bound to the
column was assessed by subtracting the total paraquat content of the
washings from the paraquat content of the original solution. The
coupling solution had been shown to contain 91.6mg paraquat
propionate by the above assay before mixing with the Sepharose.
Initial paraquat content of coupling solution 91.5mg
Total paraquat content of washings 62.6mg
Amount bound to 16m1 AH-Sepharose 29.0mg
54

Figure 2.1. Standard curve for the determination of paraquat
0
0
'o
r9
.O
2"S
2.0
1.5
1.0
0.5
0 100 200 300 400 500
Paraquat propionate pg/ml
55

Data provided by Pharmacia indicated that there were between 6 , and
10 pmol of available spacer groups per ml of AH-Sepharose which gave
between 90 and 150 pmol spacer groups on 15 ml Sepharose to which
were bound 72pmol of paraquat propionate (molecular weight 403).
This means that there was between 48 and 80% saturation of the
available spacer groups.
2.3.2 Affinity purification of the 321 antibody.
Spent medium (100ml) from a3 day culture of the 321 hybridoma
was recirculated through 15m1 of paraquat-AH-Sepharose packed in a
20m1 syringe. at 0.5m1/minute overnight at 40 C. The column was then
washed with PBS containing 1M NaCl (pH 7.2) at a flow rate of
lml/minute until no more protein could be detected in the effluent. The
column was then washed quickly with PBS containing 0.15M NaCl (pH
7.2) then eluted with 0.2M glycine-HCI (pH2.5) at a flow rate of 0.6
ml/minute. The protein content of each 0.5m1 fraction was estimated
by absorbance at 280nm and those fractions containing significant
amounts of protein were pooled giving a total volume of 2.5m1 and the
pH was adjusted to 7.0 by the addition of Tris buffer. The mouse IgG
content of the pooled fractions was assessed by ELISA as described in
section 2.2.5 and gave a titre of 1/32,000 at 50% of maximum
absorbance.
2.3.3 Determination of the protein content of the Pooled fractions.
The protein content of the pooled fraction was assessed using the
66

Biorad Protein assay, a modification of Bradfords dye binding assay
(Bradford 1976). A standard curve was constructed using bovine
serum albumin (BSA) standards ranging from 0 to l00pg/ml BSA (figure
2.2). The protein content of the pooled fraction was calculated to be
1.02mg/ml from the standard curve.
2.3.4 Assessment of antibody purity by SDS-volyacrylamide gel
electrophoresis.
The purity of the protein in the pooled fraction was assessed by
SDS-polyacrylamide gel electrophoresis using the discontinuous buffer
system of Laemmli (1970).
a) Reagents.
1) Acrylamide stock solution. 30% acrylamide, 0.8% bis-acrylamide
In distilled water. The solution was filtered and stored in the
dark at 4" C.
Ii) Reservoir buffer. 0.025M tris (PH 8.5), 0.05M glycine,
0.1% SDS.
iii) Sample buffer. 0.1M tris (pH 8.0), 0.75 mM EDTA. 1% SDS, 20%
glycerol. For reducing gels. 5% mercaptoethanol was added.
iv) Stacking gel buffer x4.0.6M tris-HC1 (pH 6.8). 0.4% SDS.
v) Separating gel buffer x4.1.5M tris-HC1 (pH 8.8). 0.4% SDS.
vi) Stain. 0.2% Coomassie blue (Sigma) in 95% ethanol. The stain was
mixed with an equal volume of 20% acetic acid immediately before
use.
57

Figure 2.2. Standard curve for the determination of protein
1.1
1.0
0.9
E 0'6
C
t11
0.7
C
m
0.6
0.5
0 N
ä 0.4
0.3
0.2
0.1
0
a
02468 10 12 14 16 18 20
Bovine serum albumin jug/ml
58

) Method.
A discontinuous gel 1.5mm thick was cast between two glass
plates. The separation gel (10% acrylamide) was overlaid with ai cm
depth of stacking gel (5% acrylamide) cast around a comb to form sample
application wells.
The cast gel was run in a vertical gel tank with the upper and lower
reservoirs filled with reservoir buffer. Protein was dissolved in sample
buffer by boiling for 5 minutes, and loaded onto the gel in the sample
application wells. A standard mixture comprising proteins of known
molecular weights (Sigma; SDS-6H) was also loaded. Electrophoresis
conditions were 40mA constant current for approximately 4 hours at
room temperature or until the dye front had reached the bottom of the
gel.
The gel was released from the plate assembly and stained for 1
hour. Protein bands were revealed by destaining in several changes of
10% methanol, 10% acetic acid in water. Destained gels were dried under
vacuum onto filter paper sheets. Under non-reducing conditions a single
band at 15OkD was observed and under reducing conditions single bands
at 50kD and 25kD molecular weight were observed representing heavy
and light chains (Plate 2.1).
69

Plate 2.1. Analysis of the purity of the 321 antibody standard
polyacrylamide gel electrophoresis
The plate shows the result of polyacrylamide gel electrophoresis of the
321 antibody standard under reducing (lane 3) and non-reducing
conditions (lanes 4 and 5). Lane 2 shows spent medium from the 321
hybridoma before affinity chromatography. Lane 1 shows molecular
weight standard marker mixture (Sigma SDS-6H) comprising
ß-galactosidase (mw 116kD), phosphorylase b (mw 97kD), bovine
albumin (mw 66kD). egg albumin (mw 45kD) and carbonic anhydrase (mw
29kD).
60

Plate 2.1
kD
116
97 -
69-
45-%
29-ýº
123
me
r
61
45
rrv -H+L
-H
-L

2.3.5 Assessment of antibody purity by high performance liquid
chromatography.
The purity of the antibody was also assessed by high
performance liquid chromatography (HPLC). A TSK 4000 column was
equilibrated in O. 1M sodium phosphate buffer pH 6.8. Samples were
injected through a 20p1 loop at a flow rate of lml/minute and
absorbance at 280nm was measured with a Beckman detector and
recorded by a BBC microcomputer.
Figure 2.2 shows absorbance traces
of the purified protein sample and the protein standards (molecular
weight markers for gel filtration; Sigma MW-GF-200) used to calibrate
the column. Integration of the relative peak areas showed that the
l5OkD peak represented 90% of the protein.
62

--. -. -. -- _, mmß
Figure 2.3. HPLC analysis of the purified 321 antibody
20111 samples containing approximately 10pg of protein were loaded In
each case. Absorbance profiles and retention times (Rt) are shown for
the purified 321 antibody standard and for standard protein solutions
of known molecular weight. The detector sensitivity was set to a full
scale deflection of 0.1 absorbance units.
63

Figure 2.3
Purified 321 antibody
JRt75s
f3-amylase
--k -
Alcohol dehydrogenase
Bovine serum albumin
Carbonic anhydrase
J\R675s
1200s
64
Rt=545s
Rt=575s
Rt=603s

Chapter 3
The Automated Assay System
3.1 Requirements for an automated assay system.
The optimisation of a cell culture system requires that regular and
precise measurements of process variables can be made and that these
measurements can be applied to control of the process. Whilst
parameters such as pH and oxygen concentrations can be readily
measured and controlled in bioreactors (Harris and Spier, 1985), the
detection and quantification of proteins and other macromolecular
products requires off-line assays which often precludes the direct
control of their production. The initial aim of this project therefore, was
to develop a system for the automated on-line quantification of
monoclonal antibody concentration In bloreactors and to apply the
system to the identification of parameters which Influence antibody
production by hybridomas.
In order to be effective as a monitoring system for antibody
production in fermenters the system must fulfil three main
requirements.
i) Sensitivity. The assay must be sufficiently sensitive to detect small
changes in antibody concentration over the concentration range seen in
a typical culture. This may vary from 1-1000pg/ml depending on the
hybridoma line and the process intensity.
ii) Reproducibility. The assay should give accurate and reproducible
results because it will not be possible to replicate samples in a
dynamic process.
65

iii) Reliability. The monitoring of product formation in bloreactors
requires that the automated assay will run reliably for several days or
weeks.
3.1.1 Methods for the monitoring of product formation in bloreactors.
High performance liquid chromatography (HPLC) and fast protein
liquid chromatography (FPLC) are both fast and reproducible methods
-for the separation and quantification of protein and because they
work on a flow through basis they can be readily automated. The main
disadvantages of such systems are that they are expensive for use In
a dedicated monitoring system and the identity of the protein is
assumed by its retention time and may be confused with molecular
species with similar retention times.
A direct sensing device would represent the ideal situation whereby
the device would generate a detectable signal only on contact with
the desired molecule. In reality the device has to be sensitised with a
highly specific mediator which will convert the analyte into a product
whose activity can be determined by more conventional sensors . such
as ion selective probes or conductimetric, thermometric or light
sensitive devices.
Enzymes have received considerable attention as mediators due to
the specificity of their catalytic activity and the ability of the
appropriate enzyme to generate detectable moieties from its interaction
with the analyte. The most common example of this type of sensor is the
enzyme electrode where enzyme is bound directly to the porous
membrane of an electrode. The probe is selective for ions generated
66

y the specific interaction of the enzyme with its substrate and because
the reaction takes place in the immediate vicinity of the probe it is a
highly sensitive system. A typical example of this type of probe is an
enzyme pH electrode to detect glucose by immobilised glucose
oxidase (Nilsson et al., 1972).
However the use of enzymes is limited because it is necessary to
have a well . characterised enzyme-substrate interaction which will
also elicit a detectable response. Such enzymes are readily available
for simple molecules such as glucose or amino acids but more
complex or novel molecules present considerable difficulties in the
Identification and characterisation of an enzyme system.
An alternative to the enzyme is the use of specific antibodies
to capture the analyte. This in itself will not generate a detectable
response for conventional probes, although research is progressing
into the use of direct sensing devices such as sensitised- piezo
electric crystals (Roedere and Bastiaans, 1983) or antibodies bound
to fibre optics (Sutherland et al., 1984) which will directly detect
antibody antigen interactions. It is therefore necessary to introduce
a second component into the assay. This, could be either labelled
antigen in a competitive assay or labelled second antibody in a
sandwich assay. The labels could be isotopic, fluorescent or enzymic
in nature making the assay analogous to the solid phase immunoassay
which is widely used in research and clinical applications. The main
limitation with conventional solid-phase assays is In the time
required for the assay due to the small available surface area and
long diffusion distances encountered in the tubes or microwell plates
which are generally used. The use of a suspension of small
particles as the solid phase in ELISA provides a much greater surface
67

area for a given volume of fluid and allows more rapid interactions
between free and surface-bound molecules. Sepharose appears well
suited as a solid phase with a small particle size of less than 100um and
a well characterised surface chemistry allowing the covalent coupling
of proteins. Sepharose beads formed the basis of the ELISA devised by
Engvall and Perlman (1971). Washing stages however required
the separation of beads from the suspension at each stage making the
process difficult to automate although an automated ELISA using
membrane separation of the beads during washing has been developed
(Ismail et al., 1981).
The use of a packed bed of Sepharose would eliminate the need
for the separation of Sepharose from the reactants in that the bed
could be perfused sequentially with the reactants and wash buffers
and would facilitate automation. Such systems have been used for the
assay of protein molecules, for example Mattfasson et al. (1977) used
anti-human albumin antibody bound to Sepharose in a competitive
ELISA for human albumin. Bound enzyme-antibody conjugate was
quantified using a thermistor to detect the heat generated from the
enzyme reaction with a substrate. They were able to repeatedly
regenerate the bed (>30 times) in order to assay further samples.
In addition Preston and Spier (unpublished results) demonstrated
the feasibility of a sandwich ELISA with Sepharose-bound antibody
for the detection of foot and mouth disease virus antigen on a
continuous basis.
The initial part of this investigation was to modify and develop this
type of assay with the aim of producing a fully automated ELISA for
the detection of monoclonal antibody in culture. For the purposes of
the assay a colorimetric determination of enzyme activity was adopted
68

as a simple detection system which could be readily automated. In the
system of Preston and Spier peroxidase-labelled second antibody was
used to develop the chromogen whereas in this study the use of
alkaline phosphatase was adopted due to the greater stability of the
substrate (p-nitrophenyl phosphate).
3.2 Immunosorbent preparation.
3.2.1 Preparation of IgG fraction from sheep serum.
Sheep anti-mouse immunoglobulin was obtained from the
Scottish Antibody Production Unit. The IgG fraction was prepared by
anion exchange chromatography as follows.
Washed DE52 cellulose (Whatman) was poured into a 2.6 cm
diameter glass column to a bed height of 30cm. The column was then
equilibrated overnight against 0.0175M sodium phosphate buffer
pH6.4. A lOml aliquot of sheep anti-mouse serum was dialysed
against the same buffer overnight. The serum was then centrifuged at
1800xg for 10 minutes to remove any precipitate and the clarified serum
was pipetted directly onto the surface of the column bed. The bed was
then eluted with sodium phosphate buffer at a flow rate of
iml/minute and 2m1 fractions were collected. The, absorbance at
280nm of the column effluent was monitored using a imm pathlength
flow cell (LKB) in an LKB Ultrospec spectrophotometer. The data was
recorded using a BBC microcomputer connected - via the RS232
port of the spectrophotometer: The absorbance of a 1mg/ml solution
of IgG was taken as 0.14 for a 1mm pathlength flowcell. Under these
conditions the IgG fraction of the serum eluted at the void volume
69

of the column whilst other serum proteins were retained. Bound
proteins were desorbed with 0.4M sodium phosphate buffer (pH 6.8) and
the column was then re-equilibrated in 0.0175M phosphate buffer as
above.
The mouse immunoglobulin-specific titre of each fraction was
assessed by competitive ELISA. Each fraction was titrated against
mouse IgG with alkaline phosphatase-labelled anti-mouse IgG as the
competing antibody. Figure 3.1 shows the recorded absorbance and the
mouse immunoglobulin specific antibody titre of the collected
fractions.
Those fractions showing anti-mouse immunoglobulin activity were
assessed for purity by SDS-polyacrylamide electrophoresis. Plate 3.1
shows a non-reducing gel of the IgG fraction from the ion exchange
column. Essentially all proteins have been removed except for a protein
band at 150kD corresponding to immunoglobulin. The bands visible at
5OkD and 25kD suggest that there was some degradation to the heavy
and light chain components of IgG.
3.2.2 Coupling antibody to Sepharose CL4B.
Immunoglobulin was bound to Sepharose CL4B by a modification of
the periodate oxidation method of Wilson and Nakane (1976).
The required amount of Sepharose was washed with distilled
deionised water to remove preservatives and fine particles. The
Washed Sepharose was added to 100 mg/ml aqueous sodium periodate
solution and mixed gently for three hours at room temperature. The
70

Figure 3.1. Purification of the IgG fraction of sheep anti-mouse serum
m
x
aJ
C-
-+--
: F-
IV
0
.c
c
Q
using anion exchange chromatography.
5
4
3
2
1
0
70 80 90 100 110
Elution volume (ml)
0.5
0.4
0.3
0.2
0.1
The figure shows the elution profile obtained from a DEAE ion exchange
column loaded with 10ml sheep serum and eluted using 1r . 5mM sodium
phosphate buffer (pH6.4).
Anti-mouse immunoglobulin titre as determined by
competitive ELISA.
Absorbance at 280nm. 0.1cm path length.
71
0
E
C
O
CO
N
GJ
Li
C
O
N
.O
a

Plate 3.1 PAGE analysis of the protein fractions obtained from the DEAE
cellulose column
The plate shows the result of PAGE analysis of fractions obtained by
anion exchange chromatography of sheep anti-mouse IgG serum. Lanes
1-5 show fractions obtained from the DEAE cellulose column
corresponding to the major protein peak as shown in figure 3.1. Lane 6
shows dialysed sheep serum before fractionation and lane 7 shows
molecular weight standard marker mixture (Sigma SDS-6H) comprising
myosin (mw 205kD), ß-galactosidase (mw 116kD), phosphorylase b (mw
97kD), bovine albumin (mw 66kD), egg albumin (mw 45kD) and carbonic
anhydrase (mw 29kD).
72

Plate 3 .1
12345 67
ýýý
6.. m -
73
+ýw
r
ýº - - 200
4-
-116
_ 97
- 69
.. -29
45

gel was then washed quickly with saline, followed by 10% aqueous
ethanediol and finally coupling buffer (0.1M carbonate buffer pH9.5).
A 1mg/ml solution of immunoglobulin in coupling buffer was added to
the activated Sepharose at a ratio of 5m1 per gram dry weight of
Sepharose and the suspension was mixed gently overnight at 40 C. After
coupling, any remaining aldehyde groups were reduced with
sodium borohydride. This also served to stabilise the Schiff bases
formed during coupling and made the matrix less prone to leakage of
antibody. The amount of antibody coupled by this procedure was
assessed' by subtracting the amount of protein remaining in solution
from the initial concentration before coupling.
The gel was finally washed with copious quantities of phosphate
buffered saline (PBS, pH7.2) and stored in PBS+0.01 % thimerosal
(Sigma) until required.
3.2.3 Assessment of the Immunosorbent In the ELISA.
Conditioned medium from a 48 hour culture of the 321 hybridoma was
used as a source of mononclonal antibody. The medium was assayed by
conventional ELISA (see section 2.2.5) using purified 321 antibody as a
standard and was found to contain 100pg/ml of IgG. The medium was
stored at- -2011 C in 5m1 aliquots and was used throughout the
characterisation of the immunosorbent.
Conditioned medium was diluted from neat to 1/10.000 in fresh
growth medium. A 60p1 aliquot of packed immunosorbent beads was
added to 1.5m1 centrifuge tubes containing iml aliquots of each
medium dilution and a medium control of fresh medium only. The tubes
74

were mixed by rolling for 1 hour at room temperature. The
beads were washed twice with Iml aliquots of PBS containing 0.1%
Tween20. An appropriate dilution of alkaline phosphatase labelled
sheep anti-mouse IgG In PBS was added to each tube and mixed for a
further hour. The beads were washed again with PBS Tween then 0.5m1
of a 1mg/ml solution of p-nitrophenyl phosphate in 0.1M carbonate
buffer pH9.5 was added to each tube. The tubes were then mixed for
30 minutes to allow the colour to develop. After centrifuging, 100p1
samples from each tube were transferred to a microtitre plate and the
absorbance at 405nm read with a Dynatech plate reader.
To ensure that the effects were not due to nonspecific
interaction on the Sepharose two control experiments were also set
up. The first control was normal sheep serum bound to Sepharose by
the same method and the second was similarly treated Sepharose but
with no bound protein. Figure'3.2 shows absorbance readings obtained
for each experiment.
It was apparent that there was a significant increase In
absorbance with increasing monoclonal antibody concentration for the
sheep anti-mouse immunosorbent. The assay was sensitive to a
concentration of 100ng/ml of antibody but the absorbance fell again
at high concentration indicating that the column had become saturated
with mouse monoclonal antibody. The controls showed no increase in
absorbance with increasing monoclonal antibody concentration
indicating that there was little non-specific interaction with the
Sepharose.
76

Figure 3.2. ELISA using anti-mouse IgG bound to Sepharose as the solid
phase
1.4
1.2
p 1.0
0
a
v
C
0.8
0.6
ä 0.4
0.2
0
0.001 0.01 0.1 1 10 100
321 Antibody (, ug/ml)
The figure shows the absorbance obtained at 410nm for diluted medium
samples containing 0.001-100pg 321 antibody/ml.
0 Sepharose with immobilised anti-mouse IgG.
Sepharose with immobilised normal sheep serum.
0 Sepharose without bound protein.
76

3.2.4 Continuous flow ELISA.
The use of Sepharose with covalently bound sheep anti-mouse
Immunoglobulin for ELISA has been demonstrated in tubes and it was
necessary to demonstrate that the ELISA could be run on a continuous
flow principle using a packed bed of immunosorbent beads.
Initial experiments were carried out with a 3mm id x 100mm glass
column (Omnifit Ltd) with a bed volume of 600p1. The column was filled
with immunosorbent and equilibrated with phosphate buffered saline
(pH7.2). The flow rate through the column was maintained at
0.25m1/minute by a Gilson peristaltic pump. Samples were injected onto
the column using a iml loop accessed via a 3-way valve (figure 3.3) and
were washed through with 2m1 PBS. Alkaline phosphatase labelled
anti-mouse IgG was then introduced through the sample loop
followed by a further wash of PBS. ' Finally p-nitrophenyl
phosphate was injected through the sample loop and allowed to react
on the column for 10 minutes. The reacted substrate was then passed
through a flow cell (1mm pathlength) in an LKB Ultrospec
spectrophotometer and absorbance readings were recorded at 10
second intervals. In the initial experiments the column was
regenerated with a 2ml wash with 3M sodium thiocyanate between
samples.
Dilutions of conditioned medium samples were assayed using the
continuous flow ELISA. Figure 3.4 shows a typical set of data from
the ELISA and figure 3.5 shows peak area plotted against antibody
concentration for the same data. The assay was sensitive to
concentrations of 321 antibody between 1 and 100 jig/ml and although
77

Figure 3.3 The continuous flow ELISA system
Spectrophotometer
78
Buffer
lp

Figure 3.4. Absorbance traces for the continuous flow ELISA
The absorbance traces were obtained using the continuous flow ELISA
for the assay of diluted medium samples containing a)1001ig, b)1011g,
c)111g, d)O. lpg 321 antibody and e) a medium control without antibody.
The assay was performed using a 3mm ID glass column containing 600u1
Sepharose with 4.5mg Immobilised anti-mouse IgG/g Sepharose.
79

E0
C
L, f1
O
0.2
0.1
0.1
0
0.1
0
Figure 3.4
0
.123 -0
123.4
Eluent volume (ml)
-
80

Figure 3.5. Assay of medium samples using the continuous flow ELISA
1.6
1.4-
E 1.2
C
I:::
1.0
< 0.4
0.2
0
0 0.1 1.0 10 100
321 antibody (jig Iml)
The ELISA curve was obtained using the continuous flow ELISA for
diluted medium samples containing 0-100lig/ml 321 antibody. The assay
was performed using a 3mm ID glass column containing 600µl Sepharose
with 4.5mg Immobilised anti-mouse IgG/g Sepharose. Each point
represents the area of the substrate peak measured for each sample.
81

the background absorbance was considerably higher than that found
with the tube assay it served to demonstrate that the assay could be
performed on the column.
3.3 The automated assay.
Having demonstrated the feasibility of a continuous flow ELISA
it was possible to develop a fully automated system operated by
computer. A BBC microcomputer was adopted for operation of the
assay because the BBC lends itself well to the input/output operations
required for the assay and because software for the control of the
LKB spectrophotometer had already been developed and could be
readily modified for the purposes of the automated assay (see section
3.3.1).
The system designed (figure 3.6) utilised 12v miniature solenoid
valves (Lee Engineering) to actuate the various additions of
reagents and a peristaltic pump (Gilson) to determine the flow rate.
A sample loop ensured that a constant volume of sample was added in
each assay cycle whilst all other reagents were added by a timed
sequence at constant flow rate.
The solenoid valves were operated by the computer via an
interface which provides a 12 volt power supply in response to a +b
volt signal from the computer. The interface was designed and built by
Karel Newman, Department of Microbiology. University of Surrey (see
Appendix I).
82

Figure 3.6. Schematic diagram of the automated continuous flow ELISA
system
Key to numbered parts.
1) Reservoir containing phosphate buffered saline.
2) Sample loop.
3) Reservoir containing alkaline phosphatase-labelled anti-mouse IgG.
4) Reservoir containing p-nitrophenyl phosphate solution (lmg; ml).
5) Reservoir containing the column regenerating buffer (see text for
details).
6) Immunosorbent column containing Sepharose with immobilised
anti-mouse IgG.
7) Spectrophotometer.
8) BBC microcomputer.
9) RS232 interface.
10) Sample injection point.
83

%D
t4nrl
d
all
m
4-- 0)
N
. i-
t0 N
3 vu
84
N
4-
H
3

3.3.1 Computer operation of the assay.
The BBC microcomputer was programmed in BBC basic. The
program is listed in Appendix II. The timed control of the solenoid
valves was made by reference to the computers internal clock. The
clock was set to zero at the beginning of the assay and was referred
to constantly to provide a record of elapsed time. The solenoid
valves were operated via the 8 bit input/output port (user port).
The user port is memory mapped by the 6622 VIA (versatile Interface
adapter) and switching of any of the 8 lines was achieved by writing
to the user port memory area designated SHEILA. This is achieved
using an operating system byte (OSBYTE) call with the A register
containing &FFF4 (write to memory) and the x and y registers
containing the address and the byte to be written respectively. Each
of the 8 output lines operated one valve so all of the 8 valves were
operated independently.
The computer also controlled the LKB spectrophotometer via the
RS232 serial data interface and data was sent back to the computer by
the same route. When coloured substrate was passed through the flow
cell in the spectrophotometer the absorbance was recorded by the
computer at 5 second intervals. This data was stored until all
readings had been taken then the computer calculated the area of the
substrate peak and stored the result on disc with the time of the
reading. The VDU screen displayed a graph of substrate peak area
against time which was updated each time an assay cycle was
completed. Because all data was stored on disc it was possible to
retrieve and display data from previous runs.
85

3.3.2 Sequence of events during the automated assay.
1) Sample was drawn from the sample reservoir into a lml sample
loop. Buffer already in the loop was run to waste until the
loop contained the sample only.
ii) The sample was run onto the Immunosorbent column followed by
PBS which flushed the sample loop. The column was washed with
3m1 of PBS to remove any unbound material.
Mouse monoclonal
antibody in the sample was bound by the immunosorbent column.
iii) Alkaline phosphatase labelled anti-mouse IgG was then run
onto the column. Anti-mouse IgG will bind to any mouse
monoclonal antibody on the column.
iv) The column was washed with a further 3m1 of PBS.
v) Enzyme substrate (p-nitrophenyl phosphate) was then run onto
the column.
vi) The flow through the column was stopped and the substrate
allowed to react for 10 minutes.
vii) Reacted substrate was washed off the column through a flow
cell in the LKB spectrophotometer.
viii) The column was regenerated by eluting the antibody complex.
ix) The column was washed with 3m1 PBS before the next sample was
applied.
The entire assay was carried out at a flow rate of
0.3m1/minute. The valve sequence was designed so that the
peristaltic pump was operating continuously eliminating the need for
a second interface to control the pump. This also alleviated power
86

spikes caused by the switching of high Inductive loads such as pumps.
The RS232 interface had been found to be sensitive to power spikes on
previous occasions but in the configuration described there were no
instances of program failure.
3.4 Optimisation of the automated assay.
3.4.1 Determination of the optimum labelled antibody dilution.
The optimum antibody dilution for alkaline phosphatase labelled
anti-mouse IgG was determined by comparing absorbance values for
medium containing 100ug/ml 321 antibody with fresh medium
containing no antibody for a range of dilutions of labelled antibody.
The optimum dilution was considered to be that which produced the
greatest ratio between background and sample readings.
Figure 3.7 shows background absorbance levels for labelled
antibody dilutions between 1/250 and 1/2000. Labelled antibody
dilutions of 1/750 and 1/1000 gave the lowest background level and
1/1000 was used as the optimum dilution in subsequent experiments.
3.4.2 Buffer composition.
The high background absorbance observed In the automated ELISA
was initially thought to be due to insufficient washing of the
immunosorbent. Increasing washing times with PBS between the various
additions of reagents did not significantly reduce the background
level. The inclusion of 0.1% Tween 20 in the wash buffer reduced
the background absorbance only slightly. The use of high ionic
87

Figure 3.7. Determination of the optimal enzyme-labelled second
C
M
O
L
antibody dilution
100
75
Li 50
m
ol
0
25
0
The assay was performed using a 6mm ID glass column containing 500{1
Sepharose with 4.5mg immobilised anti-mouse IgG/g dry weight
Sepharose. The bars represent the substrate-peak area measured at
405nm for fresh medium without antibody expressed as a percentage of
the peak area obtained for a sample containing 100ug 321 antibody. The
background absorbance levels were compared for a range of dilutions of
the enzyme-labelled second antibody.
250 500 750 1000 2000
Labelled antibody 'dilution
88

strength buffers with 0.5M salt reduced overall absorbance readings
but the ratio between background absorbance and sample readings was
not greatly increased.
It appeared that the nature of the background seen In the column
ELISA was not due to non-specific binding of antibody nor was it due
to insufficient washing of the Immunosorbent. It had been shown
previously that there was little non-specific interaction with
Sepharose in the tube experiments (Section 3.2.3). It was observed
however that the background level in the first assay cycle was
considerably lower than in subsequent cycles. This would suggest
that one or more of the assay components was not being fully eluted
from the column by the use of 3M sodium thiocyanate. In order to
determine which of the assay components was not being fully eluted a
fresh immunosorbent column was used.
Absorbance values for the following reagent combinations were
recorded on two successive assay cycles.
1. Substrate only.
2. Labelled antibody and substrate.
3.100pg 321 antibody, labelled antibody and substrate.
Figure 3.8 shows absorbance values for each step.
The background due to substrate alone did not increase on the
second cycle indicating that the labelled antibody was being fully
eluted. However the background due to labelled antibody + substrate
was considerably higher on the second cycle than on the first implying
that the 321 monoclonal antibody was not being completely eluted
from the column. The sample absorbance was similar on successive
cycles.
89

Figure 3.8. Determination of the nature of the high background levels
observed in the continuous-flow
ELISA
The assay was performed using a 6mm ID glass column containing 500it1
Sepharose with 4.5mg immobilised anti-mouse IgG/g dry weight
Sepharose. The bars represent the substrate peak area measured at
405nm. The background levels for freshly prepared Sepharose (bars A.
B and C) were compared with the background levels for Sepharose In the
second assay cycle (D, E and F).
A and D Substrate only
B and E Enzyme labelled second antibody + substrate
C and F 100pg 321 antibody + enzyme labelled second antibody +
substrate.
90

0
E 3
cv
a)
m
Y
f0
GJ
a.
a 1.
Figure 3.8
A B C 0 E F
91
1

The computer program was modified to determine whether increased
elution time resulted, in more antibody being eluted from the
immunosorbent column. The percentage background was assessed for
medium containing 100µg/ml 321 antibody after elution cycles of
varying duration. However it was found that increased elution time
had no effect on the amount of antibody desorbed from the column
(figure 3.9).
Other agents were investigated for their effectiveness in eluting
bound antibody from the column. Glycine-HC1 (0.2M, pH 2.5) is
commonly used for the desorption of proteins in affinity
purification but it was found to be only slightly more effective
than 3M sodium thiocyanate (figure 3.10). The use of 0.5M acetic acid
was recommended by Smith et al. (1977) as being more effective than
chaotropic agents in the elution of bound antigens whilst Janatova
and Gobel (1984) advocate the use of 1M acetic acid. The effectiveness
of acetic acid for column regeneration and its effect on column stability
were investigated. The use of 1M acetic acid was found to be most
effective in desorbing antibody (figure 3.10) reducing background
readings to 20-25% of the 100ug/ml sample reading. Furthermore, 1M
acetic acid did not have any immediate deleterious effects on the
stability of the immunosorbent because the substrate peak area for
100pg 321 antibody remained the same over a period of 24 hours (figure
3.11).
Molar acetic acid was therefore used as an eluting agent for the
automated ELISA. All other conditions were as described previously
with the exception that 0.1% Tween 20 was included in the wash
buffer to prevent any accumulation of protein on the valve surfaces.
92

Figure 3.9. The effect of increasing the volume of 3M sodium thiocyanate
used to elute bound antibody.
100
90
aj 80
u
C
70
L
O
60
50
0
L- 40
u
m 30
20
10
0
02468 10 12
Etuent volume (ml)
The assay was performed using a 6mm ID glass column containing 500µ1
Sepharose with 3.9mg immobilised anti-mouse IgG/g dry weight
Sepharose. The volume of 3M NaSCN used to elute the bound antibody
complex was Increased during successive assay cycles. The points
represent the substrate peak area measured at 405nm for fresh medium
without antibody expressed as a percentage of the peak area obtained
for a sample containing 10011g 321 antibody.
93

Figure 3.10. Assessment of the effectiveness of different elution buffers
The assay was performed using a 6mm ID glass column containing 500$11
Sepharose with 3.9mg immobilised anti-mouse IgG/g dry weight
Sepharose. The bars represent substrate peak areas recorded for medium
samples containing 100ug 321 antibody and medium samples containing
no antibody (hatched bars). Each bar is the average of two readings
obtained in successive assay cycles using the following elution buffers:
A 3M sodium thiocyanate
B 0.2M glycine/HC1 pH 2.5
C 0.5: ß1 acetic acid
D iM acetic acid
94

E
C
ýn
02
to
a
f0
GJ
0
0
Figure 3.10
ABC0
95

Figure 3.11. The effect of different elution buffers on the assay stability
The assay was performed using a 6mm ID glass column containing 5O0µ1
Sepharose with 4.5mg immobilised anti-mouse IgGig dry weight
Sepharose. Each set of data points represents the peak area at 405nm
recorded over a 24 hour period for a medium sample containing 100pg/ml
321 antibody.
Elution buffer
O 3M NaSCN
p 0.5bi acetic acid
Q 1M acetic acid
96

9
CD
4J
c0
cu
Q_
3
1
oý
0
Figure 311
5 10 15 20 25
Time -- (hours)
97

3.5 Assay performance.
To assess the usefulness of the assay for the detection of
monoclonal antibody in culture supernatant a standard curve was
constructed by assaying increasing concentrations of the 321
monoclonal antibody over a period of 36 hours.
A 100ml gradient mixer (Biorad) was used for the experiment with
the starting buffer consisting of fresh medium and the limiting
buffer consisting of conditioned medium diluted to contain 100pg/ml
of 321 antibody. As samples of constant volume were drawn from the
reservoir, the Increase in antibody concentration could be calculated.
The experiment was left to run automatically for 36 hours and
absorbance readings were plotted against concentration on
completion of the run. Figure 3.12 shows a typical antibody
concentration curve obtained with the automated ELISA. The plot of
antibody concentration against absorbance was linear over the range
of 5-100pg/ml of monoclonal antibody and analysis of the curve by
linear regression gave a correlation coefficient of 0.99.
The reproducibility of the automated assay was assessed by
repeated determinations of antibody In two samples containing 25 and
100pg/ml respectively. The samples were assayed alternately with 6
replicates for each sample.
Table 3.1 shows absorbance readings for samples of medium
containing 25 and 1001jg/ml. These readings corresponded to 28 and
92pg/ml as calculated from the standard curve above.
98

Figure 3.12. Standard curve for the continuous flow ELISA
E
c
0
It
2.5
2.0
IV 1.5
L
f0
a-
1.0
0.5
I
The assay was performed using a 6mm ID glass column containing 500µ1
Sepharose with 4.5mg immobilised anti-mouse IgG/g dry weight
Sepharose. The curve represents the substrate peak area recorded for
each sample plotted against the calculated antibody concentration for
that sample.
0 10 20 30 40 50 60 70 80 90 100
321 antibody pg /ml
99
)

Table 3.1 Reproducibility of the automated ELISA
Determination
Peak area
321 Antibody Baseline
25}ßg/ml 100}ßg/ml
1 1.21 2.06 0.75
2 1.23 2.05 0.81
3 1.20 2.03 0.78
4 1.25 1.99 0.72
5 1.23 1.99 074
6 1.30 2.08 0.74
Mean 1.24 2.03 0.76
Standard
deviation
0.036 0.037 0.035
100

3.6 The use of the automated ELISA for monitoring bioreactors.
The principle of using an immunosorbent column for automated
ELISA has been demonstrated under experimental conditions. The
reproducibility of the assay was comparable to a conventional plate
ELISA with a coefficient of variation of less than 6% for a limited
number of samples. The sensitivity of the assay was limited by time
and sample size but would detect paräquat antibody to bug/ml In
the configuration described. Greater sensitivity could be obtained by
increasing the sample size but this caused column saturation at the
higher antibody concentrations seen in late batch cultures of the 321
hybridoma.
When the assay was applied to monitoring of bloreactors however,
a number of problems were encountered.
3.6.1 Column comoresslon.
After prolonged use the column packing became compressed. This
typically, occured between 36 and 48 hours resulting in a reduction
of flow rate and a consequent loss of accuracy and reproducibility. It
was also found that the packing was Irreversible In that the
Immunosorbent could not be readily redispersed within the column.
Reversing the direction of buffer flow through the column merely caused
the compacted plug of Sepharose to move down the column without
dispersing. Several attempts were made to alleviate the compression
of the immunosorbent.
i) Immunosorbent column dimensions. The dimensions of the column were
101
7i

a compromise between the speed and sensitivity of the assay. The
immunosorbent volume was kept to a minimum to allow a rapid assay
and economical use of reagents. The sensitivity of the assay dictated
the minimum volume of immunosorbent that could be used.
By increasing the cross-sectional area of the column whilst
maintaining the same bed volume, the pressure drop through the bed
could be reduced which In turn reduces the compression of the
immunosorbent. In early experiments a 3mm ID xlOOmm column which
held 6001i1 of immunosorbent was used. The maximum flow rate that
could be achieved was 0.3m1/minute but such flow rates caused rapid
compression of the column packing and increased back pressure. In
an attempt to improve the flow characteristics of the column a 6mm ID
glass column was adopted. With 500p1 of immunosorbent the flow rate
could be maintained for longer periods of time but compression of the
Immunosorbent still occured after approximately 3 days. It was not
possible to increase the column diameter further without increasing
the volume of Immunosorbent and increasing the assay time.
11) Flow rate. Reducing the flow rate through the column would also
serve to reduce compaction of the Immunosorbent. However a
reduction in the flow rate increases the time required for the assay
of a single sample. A cycle of the assay takes 76 minutes at a flow
rate of 0.3ml/minute and the need to check baseline and standard
samples at regular intervals increases the time between samples to
approximately
2 hours.
iii) Direction of buffer flow. All reagents and buffers were routinely run
upwards through the Sepharose bed in an attempt to alleviate
102

compression due to gravity. This had very little effect at the flow
rates required for the assay as these were considerably more than the
settling velocity of the Sepharose. Reversing the direction of buffer
flow by inverting the column at regular intervals did not alleviate
the problem of column packing.
iv) Solid phase. The problems encountered were largely due to the
compressible nature of the Sepharose which was being used as the solid
phase. It would therefore appear that a more rigid and less cohesive
solid phase would overcome these problems. The use of a rigid particle
would also allow the assay to carried out more rapidly.
Preliminary experiments with silanised porous glass as a solid
phase were carried out. The activated porous glass was kindly provided
by Dr P. Kwasowski of the Department of Biochemistry. University of
Surrey. The IgG fraction of sheep anti-mouse antibody was coupled to
the porous glass by cross linking with glutaraldehyde and any
remaining groups were blocked by the addition of 1M ethanolamine.
These preliminary experiments however showed very high background
levels which masked any specific interaction between sheep anti-mouse
and 321 monoclonal antibody. The high background levels were
probably due to non-specific interactions with the modified
surface chemistry of the glass. It is possible that such non-specific
interactions could be reduced or eliminated by modification of the
surface chemistry of the glass or by the use of other blocking agents
but this would need much further investigation.
The coupling of proteins to other rigid particles such as modified
polystyrene (Neurath and Strick. 1981). nylon (Morris et al.. 1975;
Hendry and Herrmann, 1980) or various inorganic matrices
103

(Weetall, 1976) has been described and may prove useful as the solid
phase in ELISA but would need extensive investigation and
characterisation.
An alternative to solid phase immunoassays is the use of
homogeneous enzyme immunoassays. Such assays do not require
separation of reagents because the enzyme activity is directly
modulated by antibody antigen reactions. The most common form is
widely known under the trade name EMIT (Syva Corporation) and was
first developed by Rubenstein etal. (1972). This system is effective only
with smaller antigens such as drugs due to the configuration of the
assay.
A more promising technique for larger antigens such as antibody
molecules is the proximal linkage system where two monoclonal
antibodies against different epitopes on the same antigen are
labelled with different enzymes (Ngo and Lenhoff, 1981). In this
system, the product of one enzyme is the substrate of the other and
significant substrate conversion only occurs when both antibodies
react with the antigen bringing the labels close together.
Homogeneous enzyme immunoassays are rapid and readily
automated and could provide a powerful technique for the
monitoring of fermentation products. The main limitations with such
techniques is that the assay requires specifically designed reagents for
each antigen and the sensitivity
3.6.2 On-line sampling of the bloreactor.
is often low.
Sampling of the bloreactor is a major limitation in any on-line
system where small and reproducible samples are required. There are
104

two main requirements in sampling the fermenter which are firstly
avoiding contamination of the culture fluid and secondly the
generation of a sample free of cells and debris for the assay. The use
of a 0.2µm pore diameter filter would appear to fulfil both these
requirements, however membrane filters are very susceptible to
blocking which limits their long term use in fermenters. In an attempt
to overcome this problem a cross-flow filter was developed
(Appendix III). The device was designed to recirculate fluid from the
fermenter across the membrane to flush any debris from the
membrane surface. Samples would then be periodically drawn
through the membrane to be assayed In the automated ELISA. The
cross-flow filter could be operated maintenance-free for several days
using conditioned medium containing cells and cell debris but
blocking of the membrane could not be completely eliminated.
Furthermore, when applied to hybridoma cultures It was discovered
that circulation of cells and medium at flow rates sufficient to
prevent blockage of the membrane was detrimental to cell growth.
Figure 3.13 shows the growth of cells in a1 litre fermenter
(Gallenkamp) with and without medium recirculation.
The recirculation of cell-free medium from a fermenter would
eliminate cell damage but would generate the same problems
encountered with the filter device in that large numbers of relatively
small particles would have to be retained in the reactor by some
porous barrier. Spinning filters have been used to retain hybridomas in
perfusion reactors (Tolbert et al., 1985; Reuveny et al.. 1986) and could
provide a cell-free stream of medium which could then be sampled for
the automated assay. A further possible solution to this problem would
be the use, of an immobilised cell reactor. Immobilised cell reactors
105

Figure 3.13. The effect of medium recirculation on hybridoma growth
1.0
0.9
0.8
p 0.7
C)
0.6
0.5
0.4
C, 0.3
0.2
01
0
Each curve represents the number of viable cells in a 500m1 stirred
culture recorded over 100 hours either with (0) or without(Ö) medium
recirculation.
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
106

can be perfused by cell free medium which can be sampled by
tangential or cross-flow filters with high flow rates. However the main
limitation with these systems for the investigation of production
kinetics is that the measurements of cell numbers or biomass necessary
to determine specific production kinetics is impracticable in such
systems.
3.7 Conclusions. I
1) The use of a packed bed of Sepharose with covalently bound
anti-mouse antibody could be employed as a solid phase for a sandwich
ELISA. The increased surface area and reduced diffusion distances in a
packed bed greatly reduced the time for the ELISA to approximately 75
minutes/sample. The assay could be operated on a continuous flow basis
and was readily automated using a microcomputer.
ii) The reproducibility of the assay was comparable to a microplate
ELISA with a coefficient of variation between samples of 30 assay cycles)
without loss of sensitivity.
107

v) Column compaction occured due to the compressible nature of
Sepharose at high flow rates. Attempts to alleviate this problem by
changes in the bed geometry or periodically reversing the buffer flow
were unsuccessful. This limitation makes the assay impracticable for the
long term monitoring of bloreactors in its present configuration.
The development of a system based on a more rigid solid phase would
overcome this problem and allow higher flow rates and hence faster
assay times. However such a development would require extensive
investigation and it was considered that this would diverge from the
main aim of this investigation which was to investigate factors
affecting antibody production by hybridomas in culture.
vi) Difficulties were experienced in producing a cell-free sample from
the fermenter. Recirculation of medium through a cross flow filter
provided a cell free sample stream but caused a loss of cell viability at
the flow rate necessary to prevent filter blockage.
It was therefore decided adopt a more conventional ELISA system
using microwell plates to assay manually taken samples. Although the
conventional ELISA is a time consuming technique it does allow the
simultaneous testing of large numbers of samples and standards and
can be replicated to reduce the influence of assay errors.
108
ý,

Chapter 4
Hybridoma growth and antibody production kinetics in batch culture.
4.1 Introduction.
Many studies have been centred around the improvement of
hybridoma growth characteristics in culture (Low and Harbour. 1985a;
Low and Harbour, 1985b; Reuveny et a!., 1986a; Reuveny et al., 1986b;
Glacken et a!., 1986; Hu et a!., 1987). These strategies have led to
greater cell yields and improved cell survival with a consequent
increase in antibody yields. It has also become apparent that antibody
production characteristics vary depending on the hybridoma line. Merten
has separated hybridomas into two groups, low producers where the
antibody production rate falls rapidly during growth resulting In low
antibody titres (( 1 jig/ml) and high producers where the rate of antibody
production is sustained throughout the cell growth phase and also
during the stationary or decline phase. The latter type of production
kinetics suggests that monoclonal antibody production is not
growth-associated, a view supported by the results of continuous
culture studies by Birch et al. (1985b) and Miller et al. (1988a).
The effect of environmental factors on antibody production rate has
received less attention in the literature although some trends may be
discerned from various studies. Enhanced antibody production has been
noted in low serum or serum free medium (Glassy et al., 1988), reduced
oxygen concentrations (Reuveny et al., 1986a; Phillips et al.. 1987) low
pH (Miller et a!., 1988) or at low growth rates (Birch et a!., 1985b; Miller
et al., 1988a) although the underlying mechanisms have yet to be
elucidated.
109

The purpose of this investigation was therefore to determine the
effect of environmental factors on the production of antibodies by the
321 hybridoma and to investigate possible mechanisms for, some of the
effects described above.
4.2 Materials and methods.
Determinations of cell number and viability, all metabolite
determinations and media for cell growth were as stated in Chapter 2.
Any modifications made are described where relevant in the text.
4.3 Results.
4.3.1 Hybrldoma growth and antibody Production in static flasks.
The study of cell growth In static batch culture was carried out in
150cm2 tissue culture flasks. Cells were Inoculated at a density of
2x105 per ml in 100 ml RPMI 1640 supplemented with 10% newborn calf
serum. Cells were cultured at 37° C with 5% C0z /95% air. Cell number
and viability were evaluated daily and samples for monoclonal
antibody, glucose, lactate, glutamine and ammonia determinations
were taken simultaneously and stored at -20° C for later analysis.
Figure 4.1 shows a typical cell growth curve for the 321 hybridoma
in a static flask. Maximum cell density was achieved after
approximately 60 hours with a subsequent rapid loss in cell viability
leading to total cell death after 120-140 hours. There was no
apparent stationary phase with the 321 hybridoma cells. Maximum cell
densities achieved under the conditions described were typically
7-9x105/ml with a mean specific growth rate of 0.025h-1.
110

Figure 4.1. Growth of the 321 hybridoma in static flasks
The figures represent typical growth, metabolism and antibody
production profiles for the 321 hybridoma in"static culture. Cells were
cultured in 100ml RPMI 1640 supplemented with 10% newborn calf serum
in 150cm2 tissue culture flasks.
a) O Viable cells
Q Non-viable cells
0 321 antibody concentration
b)O Glucose
" Lactate
D Glutamine
Ammonia
111

0.9
0.8
0-7
0.6
0.5
x0.4
E
-. 00.3
E
E6
u
C-
03
0.2
0.1
0
10
9
8
5
4
`- 2
ü
J
0
0
Figure 4.1
b
20 40 6080 100 120 140
Time (hours)
0 20 40 60 80
Time
(hours)
112
101
100 120 140
s
1.8
y,
.O
C
Q
1-6 , N
N
1.4 E
E-
1-2 m
1.0 "ý
0.8 C- L
0.6 c
E
0.4
LM
D"2

With an initial concentration of 11mM the glucose was never
exhausted by the 321 hybridoma with 25-50% remaining in the medium
after cell growth had ceased. By contrast glutamine was exhausted by
about 60 hours after which there was no further increase in cell number.
Lactate accumulated in the medium reaching a plateau
concentration of 8-10mM shortly after the peak in cell density. In
static flasks approximately 1.5 moles of lactate were produced for each
mole of glucose consumed. Ammonia also accumulated In the medium
and reached a maximum concentration (1.5 mM) when Blutamine was
exhausted. The increase in ammonia concentration corresponded to
the decrease In glutamine concentration with approximately 1 mole of
ammonia liberated for each mole of glutamine metabolised.
Antibody was produced throughout the growth phase and it was
notable that antibody continued to accumulate in the medium after the
peak cell concentration. Furthermore the mean antibody production rate
(jig antibody/106 cells/hour) during the decline phase was higher than
during the growth phase (table 4.1) and approximately
60% of antibody
was accumulated in the latter half of the culture. Antibody titres of
160-180pg/ml were achieved in these cultures.
4.3.2 Hybridoma growth and antibody production in stirred batch
culture.
The study of cell growth in stirred batch culture was carried out in
a llitre Gallenkamp fermenter which was modififled by replacing butyl
rubber seals with silicone rubber and agitated using a large magnetic
follower mounted through the stirrer shaft. Aeration was provided by a
.1
113

a
L
. 4-
u
L u
. cý
'a
C
C
N
C
ra
1
4-
CD
4-
C
N
OJ
f9
L
C
O
-t-
u
0
4-
c
ri
L
C
ru
0
t7
T: -
-4t
a
Fý
a-
ö
: a-
u
>
O
-o C)
Co
Öy
O
rn
O
+1
m
rn
m
4 a ö
t
ü
c
d
+. u a
O
ö
Ci
+1
M
p. p. N
N
C)
41
-4-
c0
L
i-
o
-
O
O
+1
C
N
m
a
r Ln
c14
O
ö
c
a)
114
u
. ý-
c/)
C-
't
m
O
N
Lfi
in
O
ö
d
E
Cu
LL

95% air 5% COz mixture passed through the headspace at a flow rate of
50ml/minute. With an agitation rate of 100rpm and a working volume of
500m1 a KLa value of 1.19h-1 was achieved (see Appendix IV). pH was
monitored with a sterilisable pH probe (Russell).
Cells were inoculated at a density of 2x105 per ml in 500 ml RPMI
1640 supplemented with 10% newborn calf serum. Cell number and
viability were evaluated daily and samples for antibody, glucose,
lactate, glutamine and ammonia determinations were taken
simultaneously and stored at -201 C for later analysis.
Figure 4.2 shows a typical growth curve for a stirred batch culture
of the 321 hybridoma. Cell growth rates were higher in stirred culture
than in static culture and cell densities of 0.9-1.1 x106 /ml were
achieved after 40-48 hours with a mean specific growth rate of
0.035h-1. The decline of the culture was also rapid and total cell death
was seen after approximately 100 hours.
Glucose consumption and lactate production were similar to that
seen in static batch cultures. Glutamine was again completely depleted
by the end of the growth phase although cell densities achieved in
stirred culture were higher than in static culture. The pH of the culture
declined during the growth phase reaching a final pH of 6.7 which
probably reflects the large amount of lactic acid produced by these cells.
Antibody titres achieved in stirred batch culture were generally
lower than those seen in static cultures with titres of approximately 120
jig/ml. Mean antibody production rates during the growth phase were
also lower in the fermenter than in static flask cultures (table 4.1). As
before, the mean specific antibody production rate durl1flg the decline
phase was higher than during the growth phase.
116

Figure 4.2. Growth of the 321 hybridoma in stirred culture
The figures represent growth, metabolism and antibody production
profiles for the 321 hybridoma in stirred batch culture. Cells were
cultured in 500m1 RPMI 1640 + 10% newborn calf serum in a1 litre
Gallenkamp fermenter.
a)O Viable cells
" Non-viable cells
13 321 antibody' concentration 't
pH
b)O Glucose
" Lactate
Q Glutamine
Ammonia
116

1.0
0.9
0.8
0.7
0.6
° 0.5
x
E 0.4
cu
u 0.3
c
0
E6
°1
to
u
0
0
u2
LV
0.2
0.1
0
10
9
8
1
0
Figure 4.2
a
0 10 20 30 40 50 60 70 80 90 100
b
0 10 20 30 40 50 60 70 80 90 100
-
Time(hours)
117
I
pH
r7.2
7.0
6.8
6.6
120 6.4
100
-80
60 E
o,
40 :: `
>14
Icl
20
0Q
2.0
1.4 E
E-
1-2 m
1.0 E
0.8
ro
0.6 "E
0.4
0.2
0

4.3.3 The effect of Blutamine on cell growth and antibody production,
In the previous studies the onset of the decline phase
in batch
cultures of the 321 hybridoma coincided with the exhaustion of
glutamine. Antibody production rates were also higher during this
phase.
Stirred batch cultures at initial glutamine concentrations of 2,1.3,
0.6 and 0.3mM (0.3,0.2.0.1 and 0.05mg/ml) were set up in parallel.
Duran bottles (250m1) were used for 100mi cultures and were agitated
with a magnetic follower at 100rpm.
Figure 4.3 shows cell growth curves for each glutamine
concentration. Glutamine was exhausted in each case and cell growth
ceased at the same time. The initial growth rate was similar for each
glutamine concentration and the maximum cell density declined with
decreasing glutamine concentration, demonstrating the requirement for
glutamine in proliferating cells.
The mean specific uptake rate of glutamine was greatly reduced at
lower glutamine concentrations (table 4.2) which suggests that
glutamine uptake rate is influenced by the glutamine concentration in
the medium. Ammonia concentrations reached maximum levels when all
the glutamine had been utilised except for the lowest initial glutamine
concentration of 0.3mM where only small amounts of ammonia were
detectable.
The total amount of glucose utilised by cells was similar at 2,1.3
and 0.6 mM glutamine whereas the mean specific consumption rate
increased at the lower Blutamine concentrations (table 4.2). At
0.3mMglutamine, glucose consumption rate was reduced.
118

Figure 4.3. The effect of reduced glutamine concentration on hybridoma
growth and antibody production
Cells were cultured in 100m1 stirred suspension cultures In RPMI 1640
+ 10% newborn calf serum with glutamine concentrations of 2.0, (0);
1.3 (0); 0.6 (0) and 0.3 (0) mM.
The figures represent a) viable cells, b) 321 antibody concentration, c)
glutamine, d) ammonia, e) glucose and f) lactate.
119

1.0
0.9
0.8
0.7
E 0.6
'O 0.5
x
- 04
0)
u
0.3
m
> 0.2
E
4-
0.1
0
200
180
160
140
.
120
100
60
60
40
20
0
Figure 4.3
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
120
I/

2-0
1.8
1.6
1.4
1-2
E 1.0
a
tu 08
L3 0.6
0.4
0.2
0
2.0
1.8
1.6
1.4.
1.2
E 1.0
R
0-8
ä 0.6
0.4
o-2
0
(Cl
0 10 20 30 40 50 60 70 80 90 100
(d)
Time (hours)
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
121

10
9
8
E 6
N,
d
O
E6
E
°J
}5
v4
4
3
2
1
0
12
11
10
9
8
3
2
1
0
(e)
0 10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
122

C
O
LI
'D
O
C-
0
A
0
c
rv
c
ro
L
4-
rn
E
0
. n`
C
O
C
O
} t9
C
LI
u
C
ü
a
C_
E
ro
y--
V-
a
i--
G/
.D
ro
F--
i
c
O
u
ä
Cv
1
-
ü
rn
Si
:ö
m
Co N m
w Ö !
-in
f ß ' 1
I, "- O O O O
O
u cm
aJ
. a...
d
(9
. ý.
a
d
c %-
.E
rn
l7
N
n
0
%o
ö
o 0
ö ö, ö c>
} ;o LM un r-
3 , N N N N
O O O O
O ,ý
p O O O
(D
. r1
"
I. 7
=
E
I
,
123
Ln
fýI N C. O
O Ö
ni
ru
C-
ro
a
-t-
N
c
GJ
L.
rp
c..
N
L
GJ
L
E
D
Z
a;
Np
ý
C
CL.
3
0
1
on
rn
c
a
ro
cx
41

Antibody production rates for the growth phase of the culture
and overall production rates were calculated and are shown in table
4.2. The antibody production rate was , similar for initial glutamine
concentrations of 2,1.3 and . 0.6 mM but with a significant decrease
in the mean antibody production rate at the lowest glutamine
concentration.
It was also notable that the mean antibody production rate for the
growth phase of the culture was lower than the overall antibody
production rate, which was in keeping with earlier observations that
the antibody production rate increases during the decline phase of the
culture.
4.3.4 The effect of nH on cell growth and antibody production
The effect of pH on cell growth and antibody production was
investigated. The initial experiments were carried out by varying
the sodium bicarbonate concentration to give the required pH values
in equilibrium with 5% C02. The pH drifted rapidly due to the poor
buffering capacity of low concentrations of sodium bicarbonate. In
addition it has been suggested that bicarbonate is an Important
nutrient for cell growth and may affect cell growth independently of
pH (Geyer and Chang, 1958; McLimans, 1972). In order to overcome
these problems 20 mM HEPES buffer was included in the medium and
the sodium bicarbonate concentration was reduced to 5mM. To minimise
the effect of pH drift cell growth rates were evaluated overnight from
a culture of cells in exponential growth: Cells in early to mid
exponential phase were harvested by centrifugation and
124

esuspended in HEPES buffered RPMI1640 at pH values of 6.6,6.8.7.0,
7.2,7.4 and 7.8. Cell growth and antibody production rates were
determined after 18 hours incubation.
Figure 4.4 shows cell numbers and antibody concentrations
after 18 hours incubation. Whereas the optimum pH for cell growth was
at pH 7.2 (figure 4.5a) with a growth rate of 0.051 h-1 (cell doubling time
= 13.6 hours) the maximum antibody production rate was achieved at pH
6.8 (figure 4.5b). Antibody production and cell growth were depressed at
lower pH (pH 6.6) and high pH (pH 7.8).
4.3.5 The effect of serum on cell growth and antibody production.
The effect of varying newborn calf serum concentration on cell
growth was investigated. Cells were seeded at a density of 2x10'
cells/ml in 100ml RPMI1640 in a 150cm2 tissue culture flask.
Experiments were set up at 1,2.5 and 10% newborn calf serum. Cell
number and viability were recorded daily and samples were taken for
antibody determination.
Figure 4.6a shows "cell growth 'curves for each serum
concentration. As serum concentration was reduced cell growth
declined until at 1% serum there was little or no growth. There was
little loss of viability at low serum concentrations and cell
viability was maintained longer than with high serum
concentrations. It is notable that cell yields were comparable for each
concentration of serum.
Figure 4.6b shows antibody accumulation for each serum
concentration. The final antibody titres for each serum
125

Figure 4.4. The effect of pH on hybridoma growth and antibody
E
o"7
0-6
0.5
%0
ö 0.4
x
r_
ä
.E0.3
ä, 0.2
u
0.1
0
production.
Cells were cultured in RPMI 1640 with 10% newborn calf serum, 5mM
NaHCO3 and 20mM HEPES buffer. The pH wasadjusted to the values
shown by the addition of either NaOH or HCI. The bars represent the cell
number and antibody concentration. Error bars are the standard
deviation (n=4).
6.6 6.8 7.0 7.2 7.4 7.8
pH
126
50
40
30
20
10
0
Q,
M

Figure 4.5. The effect of pH on cell growth rate and antibody production
rate
0.06
,-0.05
-4- 0.04
ru 0.03
ý, 0.02
LS
0.01
0
6
cJ 4
u
03
rn
0

Figure 4.6. The effect of serum concentration on cell growth and
antibody production.
Cells were cultured in RPMI 1640 supplemented with 10% (0), 5% (13), 216
(") or 1% (0) newborn calf serum. The figures represent a) viable cell
numbers, b) antibody concentration and c) the specific antibody
production rate (APR) as calculated from the data shown In figures 4.6a
and 4.6b.
128

1.0
0-9
0.8
0.7
J 0.6
E
0 0.5
x 0.4
ý-' 0.3
0.2
01
0
Figure 4.6
(a)
0 20 40 60 80 100 -120
Time (hours)
129
140 160

E
rn
120
100
80
>- 60
0
c
40
20
0
(b)
0 20 40 60 80 100 120 140 160
130
Time (hours)

s
O
5
4
-' 3
Q1
ý' 2
a
0
(c)
0 20 40 60 80 100 120 140 160
Time (hours)
131

concentration were similar but the specific antibody production rate
(figure 4.6c) showed a general trend of increasing antibody
production rate with decreasing serum concentration.
4.3.6 The effect of excess thymidine on cell growth and antibody
production.
The emerging trend from the previous experiments is that
conditions which are suboptimum for cell growth may enhance antibody
production. There is some evidence that antibody production rate is
enhanced at low cell growth rates (Birch et al., 1985; Miller et al.. 1988).
The following experiments therefore investigate the effect of
preventing cell proliferation by blocking DNA synthesis.
, Excess thymidine acts as an inhibitor to DNA synthesis due to
inhibition of nucleoside diphosphate reductase (Xeros, 1962). The effect
of excess thymidine on the 321 hybridoma was investigated. 2mM
thymidine has been found to be optimum for Inhibition of DNA
synthesis In HeLa cells (Puck, 1964) and was found to be suitable for
this investigation.
Cells In early to mid exponential phase were harvested by
centrifugation and resuspended in RPMI 1640 containing 2mM thymidine
(Sigma). A control flask without thymidine was also set up.
Figure 4.7 shows the effect of 2mM thymidine on cell growth. The
control flask shows continued cell growth in the absence of thymidine
(figure 4.7a) whilst cells treated with thymidine stop dividing after
approximately 12 hours (figure 4.7b). A constant cell number was
maintained for approximately 40 hours before cells began to die. The
antibody titres were comparable despite the difference in cell number
132

Figure 4.7. The effect of 2mM thymidine on cell growth and antibody
production.
Cells were cultured in 100ml RPMI + 10% newborn calf serum either a)
without thymidine or b) with 2mM thymidine. The figures show cell
numbers (0, ") and antibody concentrations (p, M) for each experiment.
Solid and open symbols represent the results of two experiments.
133

1.0
0.9
0.8
0.7
° 0.6
x
E
0.5
0.4
ý; 03
0.2
0.1
0
1.0
0.9
0.8
0.7
0.6
x 0.5
E
R 0.4
LI; 0.3
0.2
0.1
0
Figure 4.7
0
200
160
0
10 20 30 40 50 60 70 80 90 100
Time (hours)
0 10 20 30 40 50 60 70 80 90 100
Time (hours)
134
60
40
20 E
Cn
100,
30 . CM,
i0
0
?0
r200
0
io
0
E
o cn
0
)
Xa
ä

ut the mean specific antibody production rate was higher in thymidine
blocked cells (5.1±0.42pg/106 cells/h) than in the control culture
(3.5±0.34pg/106 cells/h). These data indicate that not only does
antibody production continue in non proliferating cells, but that the
non-proliferative state appears to be more favourable for antibody
production.
4.3.7 The effect of glutamine concentration on thymidine-blocked cells.
Maximum antibody production rates were observed in the decline
phase of hybridoma cultures (Table 4.1). The onset of the decline phase
was characterised by the depletion of glutamine and it is therefore
possible that glutamine level affects the rate of antibody production in
hybridoma cells perhaps by repression of antibody synthesis. However
the effects of cell growth may obscure any specific effects of glutamine
limitation on antibody production. The effect of glutamine on antibody
production in non-proliferating cells was therefore investigated.
Cells in early exponential phase were blocked with 2mM thymidine
for 24 hours then harvested by centrifugation at 100xg for 5 minutes.
The cells were resuspended in RPMI containing 5,2,1.0.5,0.25 and
0.125 mM glutamine. Thymidine at 2mM was also Included in all media.
After 24 hours. cell number and antibody titre were determined.
Figure 4.8 shows cell number and antibody titre for each glutamine
concentration after 24 hours incubation. Cell viability remained high at
glutamine concentrations of 0.5 mM and above but there was a loss of
cell viability at 0.25 and 0.125 mM indicating the requirement of
non-proliferating cells for glutamine. Antibody titres were similar
135

Figure 4.8. The effect of glutamine concentration on antibody production
E
x'0.3
0.5
Q,. u
0.2
41
0-4
0.1
0
in thymidine blocked cells.
5210.5 0.25 0.125
Glutamine mmöles/l
50
40
30ý
2U
10
Cells were cultured in RPMI 1640 with 10116
newborn calf serum and 2m11
thymidine at glutamine concentrations of 0.125 - 5.0 mM. The figure
shows viable cell numbers and antibody concentrations measured after
18 hours incubation.
136
0
10
0

etween 1 and 5 mM glutamine but declined at lower glutamine
concentrations. This was reflected in the mean specific antibody
production rates (figure 4.9). These data suggest that there is no
repression of antibody synthesis by high levels of glutamine and that
glutamine is required for antibody synthesis as low levels of glutamine
were accompanied by a reduction in the rate of antibody synthesis.
4.3.8 The effect of nH on thymidine-blocked cells.
Variations in pH have a profound effect on cell growth rate (figure
4.5a) which In turn may influence antibody production. It is not possible
to determine whether the effect of pH on antibody production is due to
a direct effect on antibody synthesis or secretion. In order to eliminate
the effects of cell growth rate the effect of pH on thymidine-blocked
cells was investigated.
Cells were prepared as described in section 4.2.7. The cells were then
resuspended in HEPES buffered RPMI at pH values of 6.6,6.8,7.2,7.4
and 7.8. All media also contained 2mM thymidine. After 24 hours
incubation cell number and antibody concentration were determined for
each pH value.
Figure 4.10 shows cell number and antibody titre at each pH value.
Cell viability was not affected by pH over the range tested. Antibody
titres were highest within the pH range 7.0-7.4 with decreasing titres
below pH 7.0 and at pH 7.8. Mean antibody production rates calculated
from this data were also similar at pH 7.0-7.4 but were depressed
outside this range (figure 4.11). This is in contrast to the situation seen
in dividing cells where the optimum pH for antibody production was
found to be 6.8.
137

Figure 4.9. The effect of glutamine concentration on antibody production
t
a,
.o
0
o,
3,
cl:
ä2
6
1
0
rate in thymidine blocked cells.
Glutamine mM
The figure shows the effect of glutamine concentration on the mean
specific antibody production rate as calculated from the data shown in
figure 4.8.
012345
138

Figure 4.10. The effect of pH on antibody production In thymidine
u
C
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
blocked cells.
6.6 6.8 7.0 7.2 7.4 7.8
pH
50
40
30 E
cn
0
20 C
Cells were cultured in RPMI 1640 with 10% newborn calf serum, 2mM
thymidine, 5mM NaHCO3 and 20mM HEPES buffer. The pH was adjusted to
the values shown by the addition of NaOH or HCi. The figure shows the
viable cell numbers and antibody concentrations measured after 18
hours Incubation. Error bars are the standard deviation (n=4).
139
10
0

Figure 4.11. The effect of pH on the antibody production rate in
thymidine blocked cells.
6
5
4
3
2
1
0
pH
The figure shows the effect of different pH values on the mean specific
antibody production rate as calculated from the data shown in figure
4.10.
6.4 6.6 6.8 7.0 7.2 7.4 7.6 7.8 8.0
140
./

4.3.9 The effect of serum concentration on thymidine blocked cells
As with pH, the serum concentration has a considerable effect on cell
growth rate (figure 4.6a) which may obscure any direct effect of serum
on antibody synthesis or secretion. The effects of serum in
non-proliferating cells were therefore investigated.
Cells were prepared as described in section 4.2.7 then incubated in
RPMI with 2 mM thymidine and newborn calf serum concentrations of 0,
1,2,5,10 and 20%. Cell numbers and antibody titres were determined
after 24 hours incubation.
Figure 4.12 shows cell number and antibody titre at each serum
concentration. Cell viability was not affected by serum concentration
except in the complete absence of serum where a loss of cell viability
was observed. Antibody titres showed an increasing trend with
increasing serum concentration in contrast to the results obtained in
growing cultures. This implies that serum does not inhibit antibody
production and the enhanced antibody production rate seen at high
serum concentration (figure 4.13) suggests a possible role for serum in
the synthesis or secretion of antibody.
4.4 Discussion.
The data presented in this chapter showed that antibody was
produced continuously during cell growth but that there was an
apparent increase In the mean specific production rate during the
decline phase of the culture. This observation is consistent with the
observations of several other authors who have found similar
141

Figure 4.12. The effect of serum concentration on antibody production
0.5
D 0.4
. oý
a
u
0.3
w 0.2
.O
0.1
0
in thvmidine blocked cells.
0125 10 20
Serum %
70
60
50
40
30 0
E
rn
.O
20 Q
Cells were cultured In RPMI 1640 with 10% newborn calf serum and 2mM1
thymidine at serum concentrations of 0-20%. The figure shows cell
numbers and antibody concentrations measured after 24 hours
incubation. Error bars are the standard deviation (n=4).
142
10
0

Figure 4.13 The effect of serum concentration on the antibody
"6 r
N
GUI
u4
7
`- 3
cn
0
1
0
production rate in thymidine blocked cells.
The figure shows the effect of different serum concentrations on the ,
mean specific antibody production rate calculated from the data shown
in figure 4.12.
02468 10 12 14 16 18 20
Serum
143

production kinetics In hybridomas producing high monoclonal
antibody titres (Birch et al., 1987; Emery et al., 1987; Reuveny et al.,
1987). The explanations for this behaviour can be incorporated Into
two hypotheses. Emery et al. (1987) suggested that the increase in the
apparent production rate was due the release of accumulated
intracellular antibody from dying cells. They provided data
which showed Increased Intracellular antibody concentrations during
the latter phase of the cell culture. A similar investigation by Birch
and colleagues (1987) did not show such accumulation and they
concluded that the specific antibody production rate
increases during
the decline phase. Subsequent work (Renard et al., 1988) has linked
antibody production kinetics to the integral of the viable cell growth
curve which strongly suggests that antibody production during the
decline phase is secreted by the viable cell population and not
released from dying cells.
It therefore seems that antibody production is not growth-associated
and this is further supported by the finding in this investigation that
cells prevented from dividing by blocking with thymidine, continue to
produce antibody. The antibody production rate was also higher in the
non-dividing cells suggesting that a non-proliferative state Is more
favourable for antibody production.
It was also found that cell growth rates and cell yields were generally
higher in stirred culture than In static culture but that antibody yields
were often lower In stirred batch culture. The lower antibody yields were
due In part to the lower antibody production rates observed during the
growth phase In stirred cultures but may also have been contributed to
by the more rapid decline of the stirred batch cultures which would
144

educe the antibody yield. Various other comparisons between stirred
and stationary cultures have been made. The results of Seaver et al.,
(1984) were similar to those seen in this investigation. whereas Dodge
and Hu (1986), found that growth rates and cell yields were similar in
stirred and stationary cultures but did observe lower antibody titres in
stirred batch culture. By contrast Fazekas de St Groth (1983) found
lower growth rates but increased cell yields in stirred batch as
opposed to stationary culture. This range of observations may reflect
the behaviour of different hybridomas in stirred culture but may also be
affected by the type of vessel used for the stirred culture experiments.
The hydrodynamic environment may differ between the various types of
vessel and comparisons between the performance of hybridomas in
different vessels must therefore be treated with caution.
Studies on the metabolism of the 321 hybridoma showed that
glutamine was consumed at a high rate and was exhausted at the point
at which cell growth ceased. This suggested that glutamine was an
essential nutrient for hybridoma proliferation and it was found that
the cell density achieved in stirred suspension culture
decreased with decreasing glutamine concentration. Other studies on
the amino acid utilisation by hybridomas have shown that growth is
accompanied by the rapid utilisation of glutamine (Seaver et a1.. 1984;
Miller et al., 1988a).
The reduced Blutamine uptake rate at lower glutamine concentrations
may be evidence of a change in the metabolism of glutamine which is
supported by the reduced ammonia yield observed with low glutamine
concentrations. At low glutamine concentrations glutamine nitrogen
may be more efficiently utilised in the manufacture of cellular
components accounting for the lower uptake rate and lower ammonia
146

yields. However"in their study on human diploid cells, Glacken et
al. (1986) found that whilst the glutamine uptake rate by MDCK cells
was decreased at low glutamine concentration, the ammonia yield
remained relatively
constant.
Glucose was never exhausted by the 321 hybridoma under the
conditions described and glucose uptake ceased at the end of the
growth phase. Lactate accumulation followed glucose uptake very
closely with 1.5-2 moles of lactate formed for each mole of glucose
consumed. The rapid conversion of glucose to lactate appears to be a
characteristic of proliferating cells including hybridomas (Low and
Harbour, 1985; Adamson et al., 1983) and causes a substantial pH drop
which inhibits cell growth.
The decreasing glutamine utilisation rate observed at low glutamine
concentrations was accompanied by an increasing glucose uptake rate
although glucose uptake was reduced at the lowest glutamine
concentration.
These observations may reflect the reciprocal regulation of
glutamine and glucose metabolism which has been reported for other
cell types (Kvamme and Svenneby, 1961; Zielke et al., 1978). A similar
phenomenon has been observed in hybridoma cells by Hu et al. (1986)
where glutamine utilisation was found to be increased at low glucose
concentrations. This implies that hybridomas can switch between
glucose and glutamine utilisation depending on the availability of
each substrate. It is also possible that glycolysis is unable to
proceed in the complete absence of glutamine because In this
investigation, glucose utilisation ceased when glutamine was
exhausted. This may also explain the apparently contradictory findings
of Ardawi and Newsholme (1982) who found that glucose uptake was
146

stimulated by glutamine In proliferating lymphocytes because they
measured the uptake of each nutrient in the total absence of the other.
Antibody production rates were not affected by high
concentrations of glutamine in neither growing nor thymidine-blocked
cells but low glutamine concentrations restricted antibody production.
This is a surprising finding because antibody production continues for
some considerable time after the exhaustion of glutamine In batch
cultures. It may be the case that large pools of glutamate. aspartate
and alanine that accumulate in cells grown on glutamine (Ardawi and
Newsholme, 1983; Seaver et al., 1984; Merten et al., 1986) provide
sufficient Intermediates for protein synthesis but without stimulation
of cell division. This may be due to the effect of glutamine limitation on
the rate of purine and pyrimidine synthesis which In turn limits DNA
synthesis effectively arresting cells In the G1 phase (Fontanelle and
Henderson, 1969; Ley and Tobey, 1970; Zetterberg and Engstrom, 1981).
The 321 hybridoma showed an optimum pH for growth of 7.2 whilst
antibody production in growing cells was optimum at 6.8. This
phenomenon has been noted by Miller et al. (1988a) who found specific
antibody production rates were higher at pH 6.8 and 7.7 than at pH
7.1-7.4. A similar increase in the specific rate of antibody
production at low pH was also found in this investigation. This is
perhaps surprising as both protein synthesis and cell growth have been
found to be suppressed at low and high pH (Ceccarini and Eagle 1971)
due perhaps to effects on nutrient uptake (Barton. 1971; Birch and
Edwards, 1979; Miller et a1., 1988a). One explanation for this
behaviour arises from the observation that the relative rate of
immunoglobulin synthesis increases in myeloma cells in stationary phase
although the overall rate of protein synthesis is suppressed (Kimmel.
147

1971). This may be due to a preferential translation of immunoglobulin
mRNA in stationary or nutrient starved myeloma cells (Sonenshein and
Brawerman, 1976): This suggests that conditions which suppress cell
growth do not necessarily suppress, and indeed may favour
immunoglobulin synthesis.
A further finding in this investigation was that that the optimum pH
for antibody production in thymidine-blocked cells was different from
that seen in proliferating cells with the optimum production rate at p11
7.1-7.4. This suggests that the effect of low pH on antibody production
rate-in proliferating cells may be a reflection of the changes in the
demands on the cells energy and metabolite reservoir at low cell growth
rate. A pH range of 7.1-7.4 provides a more favourable environment for
protein synthesis as illustrated by a more rapid growth rate but in the
absence of cell division this extra capacity for protein synthesis may be
diverted to antibody production.
The concentration of newborn calf serum in the medium had a
significant effect on cell proliferation. Cell growth was supressed
at lower serum concentrations suggesting that serum contains a
component which is essential for cell proliferation. However as high cell
viability was maintained for longer periods of time in low serum
concentrations it would appear that high concentrations of the
component are not necessary for cell survival.
It was also found that specific antibody production rates in growing
hybridoma cultures were enhanced in media containing less serum. This
phenomenon has been noted by others who have found that the
specific production rate for antibody is generally higher in
serum-free or low serum media than in serum-supplemented media
(Tharakan and Chau. 1986c; Glassy et al., 1988) which has led to the
148

suggestion that factors in serum inhibit antibody production (Glassy
et al., 1987). However it has been shown in this investigation that this
is probably not the case because antibody production was found to be
enhanced at high serum concentrations with thymidine-blocked cells.
This would tend to suggest that the effects of serum on antibody
production seen in proliferating cells may be due to the effect of the
reduced cell growth rate at low serum concentrations.
This hypothesis
is supported in the review by Glassy et al (1988) where it was shown
that hybridoma growth rates in serum-free media are generally lower
than in serum-supplemented media. The increased antibody yields at
high serum concentration
in blocked cells may be due to the stimulation
of nutrient uptake by serum factors such as growth factors (James and
Bradshaw, 1984) or insulin (Schubert, 1979) providing a larger
metabolite pool for protein synthesis.
The various findings discussed above suggest that the production
of antibody by hybridomas is affected by the both the rate of cell
proliferation and the metabolic state of the cell. It appears that more
slowly growing or non-dividing cells have a higher rate of antibody
production. This may be due to a change in the demands on the cell's
metabolite pool as the rate of cell growth is reduced. At high growth
rates the rapid synthesis of structural components for cell growth and
division places a greater demand on the cell's energy or metabolite pool
at the expense of antibody synthesis.
It has been shown in this investigation that while certain conditions
which were suboptimal for growth increased the apparent rate of
antibody production in growing cells, the optimum conditions for
antibody production In thymidine-blocked cells were similar to the
optimum conditions for growth in dividing cells. This may be due to
149

enhanced cell metabolism or nutrient uptake in ideal conditions which
would normally lead to rapid cell division. In the absence of cell division
the metabolic state of the cell will greatly influence antibody
production.
The influence of cell growth rate on antibody production makes it
difficult to determine the effects on antibody production of
environmental factors which also affect the growth rate. In order to
investigate more fully the effect of cell physiology on antibody
production rate it would therefore be necessary to maintain cells at a
constant growth rate. In the next chapter therefore, the effect of cell
growth rate on antibody production and cell physiology was examined
under steady state conditions which were achieved using chemostat
culture.
150

5.1 Introduction.
Chapter 5
Continuous culture
The data generated from batch cultures in the previous chapter
suggests that factors that affect the growth rate of hybridomas may
be important in the production of monoclonal antibody. The nature of
batch cultures however means that metabolite concentrations and
physiological conditions are continuously changing making it difficult
to accurately identify those factors which influence cell productivity.
In order to investigate the effects of growth rate or environment on
cell metabolism it is desirable to maintain cells in a constant
environment. This can be achieved by the use of a chemostat.
The theory of cell growth in continuous culture is based on the
principle of Monod (1942). that at submaximal growth rates, the growth
rate of a cell is determined by the concentration of a single
growth-limiting nutrient. Monod showed that the growth of an organism
on a limiting nutrient can be described by the equation:
p= ileax S (1)
S+ Ka
where p is the growth rate, s is the substrate concentration, pm. x is
the maximum growth rate and Ks is the substrate affinity constant. This
equation is analogous to the Michaelis-Menten model of enzyme
151

kinetics with Ks equivalent to K..
The chemostat, which was first demonstrated by Monod (1960)
and Independently by Novick and Szliland (1950), is a reactor which
maintains a constant volume with a constant rate of medium addition
such that cells'and medium are removed at the same rate. The medium
feed has a single growth limiting nutrient so that' at feed rates less
than the maximum growth rate of the cell the cell growth rate is
determined by the feed rate or dilution rate (D). At a 'constant dilution
rate a steady state will be established during which the cell number
and the physiological environment are constant, and the growth rate
(p) is equal to the dilution rate. The theoretical principles of
chemostat culture have been discussed more fully by Pirt (1975) and
Herbert et al. (1956).
Attempts to grow mammalian cells in a chemostat were made as
early as 1958 by Cooper et al. but the limited success by early workers
in maintaining a stable steady state may explain why comparatively
few studies of animal cell physiology have employed the chemostat.
Most of the chemostat studies that have been made have used glucose
as a limiting substrate (Moser and Vecchio, 1967; Tovey and
Brouty-Boye, 1976, Tovey 1980) although in many cases the limiting
nutrient was not known (Griffiths and Pirt, 1967; Van Hemert et al.,
1969; Fazekas de St Groth, 1983). Tovey and Brouty-Boye (1976)
showed that it was possible to maintain long term steady states and
that the'growth of animal cells in chemostats correlates well with the
Monod equation (equation 1).
There is still little published work on the growth of hybridoma
cells in chemostats but Birch's group in the UK (Boraston et al., 1982,
Birch et al., 1986b) and Millers group (Miller et al., 1988a, b, 1989a, b)
162

in the USA have published data derived from chemostat culture. These
studies have suggested that antibody production is not directly linked
to cell growth rate and there is evidence that antibody production
declines at high growth rates.
In this study chemostat culture has been used to investigate
the effect of growth rate on monoclonal antibody production by 321
hybridoma cells. The data indicate that antibody production rate
decreases at higher growth rates. The utilisation of Blutamine and
glucose and the production of ammonia and lactate have also been
monitored and the effect of growth rate on the metabolic quotients for
these metabolites is discussed. The monitoring of these major
metabolites should provide information on the changes in cell metabolic
activity at different growth rates.
Glutamine was selected as the limiting nutrient for these
studies as the results of the previous chapter suggest that glutamine is
the first nutrient to be exhausted In batch cultures and the
stationary/decline phase begins at that time. It was therefore decided
that a glutamine-limited chemostat might provide useful information
on the production of antibody under steady state conditions with low
glutamine concentrations.
5.2 Materials and methods.
5.2.1 Apparatus for continuous culture of hybridoma cells.
The vessel used for chemostat culture was a1 litre glass
reactor with a stainless steel headplate (Gallenkamp). The temperature
in the vessel was maintained by a water jacket connected to a
163

circulating water bath at 37° C. Agitation was provided by a large
Teflon-coated stirrer bar fixed through the stirrer shaft and driven by
magnetic coupling. The agitator speed was maintained at 100rpm.
Aeration and pH control of the culture was achieved by passing
95% air 5% C02 through the headspace at 311tres/h.
The pH was monitored by a Russell autoclavable pH probe
connected via an amplifier to a BBC microcomputer. The computer was
programmed to record pH data at 30 minute intervals. The data was
displayed on screen with the elapsed time and also stored on disc so
that it could be retrieved at a later date.
Medium was fed to the reactor from a 2litre reservoir which could
be refilled via a sterile coupling. The medium reservoir was kept at
between 0 and 40C. Waste medium was collected in a 1011tre vessel
which was connected via a sterile coupling so that the vessel could be
replaced when full. The medium flow rate was controlled by a
peristaltic pump (Gilson). Figure 6.1 shows the apparatus used for the
continuous culture of the 321 hybridoma.
5.2.2 Continuous culture.
For the continuous culture a 500m1 working volume was used.
The medium comprised RPMI 1640 supplemented with 2g/I sodium
hydrogen carbonate and 10% newborn calf serum with glutamine as
the limiting nutrient. The actual glutamine concentrations used are
mentioned at the relevant points In the results section. Cell numbers
and viability were determined daily and samples were also taken for
antibody, glucose, lactate, glutamine and ammonia determinations.
154

Figure 5.1. Apparatus used for the chemostat culture
Samp
port
155
Waste reservoir
it

. Sample volumes were typically 5m1 representing 1% of the culture
volume. Steady state was said to be achieved when five consecutive
samples gave comparable cell counts and metabolite concentrations.
5.2.3 Metabolic quotients.
The metabolic quotient (q) for a metabolite is the specific rate
of uptake or production of that metabolite under steady state
conditions.
If perfect mixing within the reactor is assumed such that the
concentration of a metabolite (C) in the outlet is the same as the
concentration of C within the reactor, the production rate (P) for a
metabolite could be described as follows:
P=FCout -FCin +VdCout
where F is the flow rate, Ci a is the concentration of a metabolite
in the inlet, Cout is the concentration of a metabolite in the
outlet, V is the volume of the chemostat and dCo ut /dt is the rate of
change of concentration of metabolite in the reactor.
At steady state however Co ut is constant and dCo ut /dt =0
Furthermore the concentration of a product in the inlet is also zero
therefore:
P=FCout
156
dt

The specific production rate qp is
qp = FCont
where x is the number of cells/ml
Because the dilution rate is given by D= F/V
then
Vx
qp = DCout. (2)
The rate of uptake (U) at steady state can similarly be described by:
U=FCIn -FCout,
and the specific uptake rate (qu) by:
x
qu = D(Cin - Cout). (3)
R
It should be noted that these equations hold true only if cells and
metabolites are perfectly mixed and adherence of cells to the walls of
the vessel can cause deviations from these equations. However,
adhesion of the 321 hybridoma to the vessel walls did not occur and
perfect mixing was assumed.
157

5.3 Results.
5.3.1 Establishment of steady state growth.
The vessel was seeded at a cell density of 2x105 cells/ml. The
cells were grown initially without the medium feed for 24 hours to
allow cells to attain the exponential phase. Medium containing 0.7 mM
glutamine was then introduced at a flow rate of 20.8 ml/hour
corresponding to a dilution rate of 0.042h-1. The cell number at this
point was 3x105 cells/ml with 98% viable. Figure 6.2 shows the progress
of the hybridoma during the initial stages of the continuous culture,
The cells quickly reached equilibrium at 3.6xl05 cells/ml and
measurements of the glutamine concentration showed residual levels
of 0.2 mM. Increasing the glutamine feed concentration to 1mM did not
result in an increase in cell number which Indicated that the cells were
not glutamine limited. However when the dilution rate was reduced to
0.024h-1 there was an immediate increase in cell number to 6.0x105/ml
and the residual glutamine concentration fell below detectable levels.
A glutamine pulse confirmed that cells were substrate limited because
the addition of 75mg of glutamine to the fermenter led to a transient
Increase in cell number which then returned to the previous steady
state level on depletion of the excess glutamine (figure 5.3). It was
also observed that when the dilution rate was subsequently increased
back to 0.042h-1 the cell number was maintained at the higher
level (6.0x105/ml).
168

Figure 5.2. Initiation of the chemostat culture of the 321 hybridoma
0.8
ö 0'7
0.6
X 0.5
y 0.4
0.3
C, 0.2
0.1
0
O
Q: Pooo
Q
Q
Q°
bo
O
00O
0
Q 13
Q
00 OOO0 ,0
0
0 100 200 300 400
Time (hours)
0
x
2.0 E
C,
m
1.0 LD
The figure shows the viable cell numbers and the residual Blutamine
concentrations observed during the Initial stages of the chemostat
culture. The nutrient feed comprised initially RPMI 1640,10°o newborn
calf serum and 0.7mM glutamine introduced at a dilution rate of 0.042
h-1 after 24 hours (first arrow). The glutamine feed concentration was
adjusted to 1. OmM/l after 96 hours (second arrow). The dilution rate
was reduced to 0.024h-1 after 180 hours (third arrow).
159

Figure 5.3. The effect of a glutamine pulse on the cell concentration
%0
I
O
e--
X
E
H
a
u
ö E
E
a
E
m
Z)
l_7
i
AA
- 30 =20 -10 0 10 20 30 40 50 60 70 80 90
Time (hours)
The figure shows cell numbers (0) and residual glutamine
concentrations (0) after a pulse addition of 0.26mmoles of glutamine to
the fermenter.
160

5.3.2 The effect of the specific growth rate on the metabolism of the 321
hybridoma.
Steady states were established at three specific growth rates of
0.024,0.033, and 0.042h-1 while cell washout occured at a dilution
rate of 0.05h-1. Figure 5.4 shows cell growth and metabolite
concentrations
during successive steady states. Steady state values for
cell number and glucose. lactate, Blutamine, ammonia and antibody
concentrations were determined for each growth rate as shown in
table 5.1. Steady state growth was established for each of the three
dilution rates on at least two occasions and was found to be
reproducible. The data in table 6.1 represents the mean ± standard
deviation for at least 10 data points determined during two or more
steady states for each dilution rate.
Table 5.2 shows metabolic quotients for antibody, glucose,
lactate, glutamine and ammonia for each specific growth rate.
5.3.2.1 Cell numbers and viability.
Cell numbers were comparable at each dilution rate and the cell
viability did not drop below 99%. A dilution rate of 0.05h-1 caused
loss of cells due to washout (figure 6.5). The rate of cell loss at
washout is related to the maximum growth rate (po. x) by the following
equation:
p ax = lnxt -Inxo +D (4)
t- to
161

Figure 5.4. Steady state cell and metabolite concentrations at different
growth rates
The figure shows steady state cell numbers (0) and steady state
concentrations of antibody (O), lactate (Q) and glucose (") at dilution
rates of 0.024,0.033 and 0.042 h-1. The arrows represent the times at
which the dilution rate was changed.
162

X0.5
0.8
0.7
0.6
N 0'4
0.3
120
Q 100
T 80
-0
0
-c 60
< 40
-r
E6
+- 2
fU
ö
E
U,
0
0
10
Lm 2
Figure 5.4
0
°°QQ
0.024 0.033 0.042
0 00 00 000000 ooö
0 0
°Q
Q
Q
i 00000 0 00
4 1"""
"
"
"""""
""
°
°
013
°°
000000 00
690 790 890 990 1090
"""
Time (hours)
163
""
""
°°
o00 000
°Q° J°Q
00 000 000
""
"
"" "
1190 1290

p
O
"1_
CD
-: C
a
L)
C
O
u
a
0
4-
a
E
rv
Cu
. 4-
N
tu
a
in
C
0
a)
. 4-
fU
L
c
0
'4 --
0
F--
LI
.n
co
-
rv
E E
c
ý
+-
E
m m ö#
-oI
ö 0
+I +1 +1
m
O O +I
3 o
m
--J
ö
E
o
Ln
N
O O O
+1 +I +1
CC) T
N
It
L! 1
N m
0
o
+t
+1
3 E
O
N o o
n
LO .ö %0 %0
0
E
OD O N
m cv m
-ý ` +I +I +1
a- o
Q
co
E
M
C
E
Iö'
o
ö
+I
ö
ö
+1
ö
cD
+I
w
V-
x ý
ö
Ln
o
n
Ln i
cD
' ö
0'
m
ö
.c .
ö ö ö
164
N
ö

N
a)
-F-
fU
C-
s
0
C-
rn
-1-
C
Q)
L-
a)
V--
'v
-I-
(U
N
4-
(1)
:Z
O
u
0
. 13
N
a
Lll
a)
rv
F-
(0
-C N +'t ýt
4
O
E o
E ö ö ö
a
c
N
0
CC)
En
"%O
O L p O CD
p1 O O O
0
C!
u
rv
ü ö N
m
a
N
O r-
N
A
Qj
co --t
rn
Ö Ö
O
'ej
o Ln - m
ru Z
'O
ö
ö
m
0
6
0
6
165

Figure 5.5. Cell washout kinetics at high dilution rate
1.0
0.9
0.8
0.7
0.6
X 0.5
N 0.4
C,
ý-' 0.3
0.2
0.1
0
140 1480 152 0 1560 1600 1640
The figure shows the effect on cell numbers of increasing the dilution
rate from 0.042 to 0.05h-1. The cell numbers recovered after adjusting
the dilution rate to 0.024h-1. The arrows represent the times at which
the dilution rate was changed.
Time (hours)
166

where xo is the population at time=0 and xt the population at time=t.
This gave a theoretical p ax of 0.040h-1 which is slightly lower
than the highest growth rate achieved in the chemostat. This value
is somewhat lower than the maximum growth rate achieved in batch
cultures but this may reflect the fact that the chemostat was
operating at a pH of between 7 and 7.1 which is slightly below the
optimal pH of 7.2 (Chapter 4, figure 4.5a).
5.3.2.2 Glutamine and ammonia.
Residual glutamine concentrations
in the fermenter were below the
limit of detection (

Sumbilla et al. (1981), which was of the order of 2-3x10-4 M.
5.3.2.3 Glucose and lactate.
Steady state concentrations of glucose remained relatively
constant for all three dilution rates while lactate concentration
increased approximately twofold from the lowest to the highest dilution
rate. While the metabolic quotients for glucose and lactate increased
with increasing dilution rate the ratio of qi sctat, to qo iucoss was not
constant in that the amount of lactate liberated for each molecule of
glucose consumed increased at higher growth rates.
5.3.2.4 Antibody.
The steady state concentration of antibody in the culture
decreased with increasing dilution rate and q. nt Icody in was 5.4,4.6
and 3.8pg/106cells/h at dilution rates of 0.024,0.033 and 0.042h-1
respectively.
Cell yields for glutamine were similar for each dilution rate
but product yields (mg antibody/mg glutamine) declined at higher
dilution rates (table 5.3).
5.3.3 Hybridoma stability.
The chemostat culture of the 321 hybridoma was maintained
for 112 days representing approximately 100 generations. At
comparable dilution rates and cell density the specific antibody
168

CL)
C
_
.E
4-
41-
E
0
L
4-
a
.T
-F-
u
0
L-
CL
ft,
Li
Lfl
4!
I--
. cj
a'
C_
E
°'
,0
L) 0
c
CL)
C
: " E
a ru
3
o
C-
U)
rn
cn
Q,
u
ý
.c
CO LO ro
ö 0 0
CP ap o`
nm m Cn
rn
ö
c
ö 0
ö ö ö
169

production rate was similar throughout the chemostat culture. This
implies that the production of antibody by the 321 hybridoma was
stable under continuous culture conditions which are highly selective
due to the limitation of an essential nutrient.
5.4 Discussion.
A glutamine limited chemostat culture of 321 hybridoma cells was
maintained at three different dilution rates over a period of 4 months.
Cell numbers were comparable at each dilution rate and cell viability
remained above 99% during each steady state. The high cell viability
can in part be attributed to the washout of non-viable cells and the
dilution of toxic metabolites.
Batch culture experiments had indicated a maximum growth
rate of 0.051h-1 but this could not be achieved in continuous
culture in that washout of cells occurred at a dilution rate of 0.05h-
1. Calculation of Amax by washout kinetics suggested a maximum
theoretical growth rate of 0.04h-1. This is slightly lower than growth
rates already achieved in the chemostat. This behaviour may be a
result of the increasing lactate concentrations observed at higher
growth rates. This in Itself might suppress growth rates by inhibition
(Glacken et al., 1986) and it has indeed been demonstrated that lactate
inhibits the growth of the 321 hybridoma cell independently of pH (J.
S. Thorpe. personal communication). However it is also likely that the
sub optimal pH in the chemostat is suppressing cell growth. It has
been demonstrated in the previous chapter that small deviations
from the optimum pH of 7.2 lead to a substantial decrease in the
170

growth rate of the 321 hybridoma. The data generated in the previous
chapter (figure 4.5a) suggest that at pH 7.0 the maximum specific growth
rate would be 0.04h-1 which agrees with the growth rate calculated by
washout kinetics.
The cell yield from glutamine observed in chemostat cultures
was similar for each dilution rate and was . higher than achieved in
batch cultures. The observation that glutamine limitation results in
increased cell yields agrees with the findings in batch cultures where
the lowest initial glutamine concentrations gave the highest cell yields
(figure 4.3).
In batch cultures with high Initial glutamine concentrations,
as demonstrated in the previous chapter, there was typically 1 mole
of ammonia liberated for each mole of glutamine consumed whereas In
the chemostat the ratio was 0.5 at p=0.024h-1 and decreased with
increasing dilution rate to 0.3 at p=0.042h-'. These observations
suggest that the route for the metabolism of glutamine may be
influenced by glutamine concentration. At low glutamine
concentrations it is likely that there is more efficient incorporation of
glutamine nitrogen into biomass. At high glutamine
concentrations, increased ammonia concentrations and lower cell
yields suggest that there is greater oxidative degradation of
glutamine. This Is in agreement with the findings of Glacken et al.
(1986) who found that maintaining low glutamine concentrations by
periodic feeding greatly reduced the ammonia yield.
Greater utilisation of exogenous glutamate under glutamine
limitation would reduce the amount of ammonia liberated due less
degradation of glutamine. It has been suggested that the
171

glutamate: 2-oxocarboxylate aminotransferases are more important
than glutamate dehydrogenase in the conversion of glutamate to
a-ketogiutarate in proliferating lymphocytes (Ardawl and Newsholme,
1982) implying that ammonia is primarily a result of glutaminase
activity. It has also been suggested that exogenous glutamate is
primarily transaminated to a-ketoglutarate whereas glutamate
derived from glutamine is oxidatively deaminated by glutamate
dehydrogenase (Borst, 1962; Kovacevic, 1971; Schoolwerth and
LaNoue, 1980). This may be an effect of compartmentation
because both
glutaminase and glutamate dehydrogenase are located within the
mitochondrion 'while the transaminases are located predominantly in
the cytosol.
There is also evidence to suggest that the flux through the
metabolic pathways used to catabolise glucose and glutamine change
with increasing dilution rate. As the dilution rate is increased the ratio
of glactate: qglucoaa increases 'while the ratio of
qa o n1a : qo 1utamine decreases. A number of explanations for this
behaviour could be put forward.
i) More glucose is converted to lactate at higher growth rates. It has
been noted that in batch cultures of hybridomas the number of moles
of lactate formed from each mole of glucose Increases with growth rate
(Luan et al.. 1987; Hu et a1.. 1987). It also appears that as the rate of
glucose uptake is increased carbohydrate molecules are Increasingly
shunted through glycolysis to lactate (Zielke et a!.. 1978). At slower
rates of glucose uptake or with more slowly metabolised sugars such
as galactose or fructose it has been suggested that more carbohydrate
172

is shunted through the pentose phosphate cycle with the formation of
less lactate (Reitzer et al., 1979). The data generated in this
investigation would support such a hypothesis.
In his studies on mouse LS cells in chemostat culture, Sinclair 0 980)
found that while the amount of glucose converted to lactate was
variable, the amount of glucose oxidised was constant over a wide range
of growth rates. Based on his calculation that the metabolic quotient for
glucose oxidised = qo 1acos e- 0.5(gi act. to), a similar constant value
is obtained at the 3 growth rates studied in this investigation (qg iacoae
0.5(gi actat e) = 0.14-0.16 )imol/106 cells/h). This suggests that at a
glucose uptake rate of 0.14-0.16 pmol/106 cells/h there should be no net
yield of lactate but such a prediction was not tested.
Ii) More glutamine is incompletely oxidised to lactate at higher growth
rates. Zielke and colleagues (1980) discovered that 13% of glutamine
carbon was converted to lactate in HeLa cells but the contribution to
qi act ate from glutamine might be expected to be small and it is likely
that much of the increase in gi. ct. t. is attributable to increased
glycolytic activity.
iii) Since mammalian cells utilise glutamlne in the synthesis of
asparagine, purines, pyrimidines and amino sugars (Levintow et al.,
1957; Salzman et a!., 1957; Meister, 1962) it is likely that the.
requirements for glutamine increase in more rapidly growing
cells. The metabolic quotient for glutamine does indeed increase with
dilution rate in the chemostat whereas q... o. i. was similar at all three
growth rates. The decreasing ammonia yield from glutamine in the
173

chemostat suggests that a greater proportion of glutamine nitrogen is
being incorporated into cell biomass. However this investigation has
also demonstrated that the cell yield from glutamine is constant over
the range of dilution rates which may be evidence that glutamine carbon
is accumulating in amino acid pools such as glutamate or alanine (Seaver
et al.. 1984; Merten, 1986; Miller et al.. 1988a).
iv) There is evidence that an increasing flux through the glycolytic
pathway inhibits glutamine oxidation (Zielke et al., 1978; Hu et al.,
1987). The reduced yield of ammonia from glutamine seen in this
investigation would be consistent with a reduction in the oxidative
deamination of glutamine. The mechanisms of such regulation have not
been elucidated but it is known that increased rates of glycolysis inhibit
oxidation via the Krebs cycle (Frame and Hu, 1984; Glacken et al., 1986)
possibly due to an increased cytosolic ATP/ADP ratio (Miller et al.,
1988a). Recently Glacken (1988) has postulated that the increased rate
of ATP production from glycolysis depletes the intracellular phosphate
pool which leads to a reduction in phosphate-regulated glutaminase
activity.
Antibody production showed a decreasing specific production
rate and product yield from glutamine, with increasing dilution rate.
This is consistent with other observations on the production of
monoclonal antibodies 'by hybridomas. Higher specific production
rates have been noted in perfusion culture systems with low cell
growth rates (Van Wezel et al., 1985; Reuveny et al., 1986b) and Miller
and colleagues (1988a) showed a similar relationship between growth
rate and the specific production rate of antibody in a glucose limited
174

chemostat culture. Birch and colleagues (1985b) also made similar
observations in chemostat cultures of hybridomas under oxygen or
glucose limitation but contrary to the findings in this investigation
they found that antibody production was depressed at low growth
rates under glutamine limitation. The reasons for this difference are
unclear and a lack of data concerning the culture conditions under
which Birch and colleagues made these observations make it difficult
to identify possible contributing factors.
In general it appears that low growth rates favour higher
antibody production rates in hybridomas. A number of hypotheses
could be put forward to explain this behaviour.
1) The capacity for macromolecular synthesis by hybridomas Is finite
and monoclonal antibodies take up a substantial proportion of the
cells capacity for protein synthesis. Hybridomas may produce in excess
of 1000 molecules immunoglobulin/cell/second (Merten, 1987). It would
therefore be reasonable to assume that at higher growth rates more
of the cells synthetic capacity would be diverted towards
macromolecular synthesis for the processes of cell division at the
expense of antibody synthesis. The changing fate of glucose and
glutamine at higher growth rates may be evidence of this change in the
metabolic balance. The Increased efficiency of nitrogen incorporation at
high growth rates may reflect the increasing demands on the cell
metabolite pool while the increasing conversion of glucose to lactate
could be supplying the increased energy needs of the cell.
ii) Antibody synthesis is repressed or inhibited by a metabolite. If
antibody synthesis is being adversely affected by a metabolite such as
176

glucose, glutamine, lactate or ammonia, then increased rates of uptake
or production at high growth rates could account for the decline in the
rate of antibody synthesis.
If such a phenomenon exists in hybridomas it is a complex process
as many factors have been shown to influence antibody production.
Glucose, oxygen (Birch et al., 1985b; Miller et al., 1988a) and glutamine
(this investigation) limited chemostat cultures show decreasing
antibody production rates with increasing growth rate. There is some
evidence to suggest that these effects are not due to
inhibition/repression in that Miller et al. (1987) have noted that the
optimum oxygen concentration for cell growth was different from that
for antibody production but g91ta. init , goiuco.. and goxYoen were
maximal at the highest qa ntibodV which would not be consistent
with inhibition or repression by any of nutrients. In addition glucose
does not affect antibody production in batch culture (Low and Harbour,
1985; Tharakan and Chau, 1986), Bodeker et al. (1988) found antibody
production by immobilised hybridomas was favoured at high oxygen
consumption rates and in this investigation high glutamine
concentrations had little effect on antibody production in batch cultures
(Chapter 4, sections 4.2.3 and 4.2.7).
Ammonia has been found to inhibit the production of viruses (Eaton
and Scala, 1961; Jensen and Liu, 1961), interferon (Ito and McLimans,
1981) and antibody production (Reuveny et a1., 1986a) but in this case
q... o nia remains unaltered at each dilution rate and steady state
ammonia concentrations decline at high growth rates. It therefore seems
unlikely that inhibition by ammonia Is responsible for the decline In
antibody production rate observed in the chemostat.
176

Increasing concentrations of lactate at higher growth rates may
Inhibit antibody production as suggested by Glacken et al. (1986).
however Reuveny et al. (1986a) and Miller et al.
(1988b) have found
that lactate levels above 20mM did not affect antibody production. It
Is also notable that the highest rates of antibody production often
occur In the latter stages of batch cultures when lactate
concentrations are highest (see Chapter 4).
iii) Cell cycle distribution. It has been reported that some
lymphoblastold cell lines produce antibody in a cell-cycle
dependent fashion with the highest production during the Gi phase
(Buell and Fahey, 1969; Takahashi et al., 1969, Byars and Kidson,
1970, Garatun-Tjeldsto et al., 1976; Turner et al., 1986). It has also
been found that the distribution of cells throughout the cell cycle
changes with growth rate where the S, Gz and M phases remain
relatively constant and variations in the length of the cell cycle can
be attributed to variations in the length of Gi. This has also been
demonstrated in chemostat culture with yeast (Bailey et al., 1982;
Scheper et al., 1988). They showed that proportionately more cells
were in the Gi phase at low growth rates. If the synthesis of
antibody by hybridomas is restricted to the Gi phase then it would
follow that the rate of antibody synthesis would be higher at low
growth rates. This could also account for the higher production rates
observed in stationary phase or in Gi arrested cells (Chapter 4). It is not
known whether immunoglobulin synthesis by hybridomas is under cell
cycle regulation but this possibility was entertained by Low and Harbour
(1985a) who detected a sub-population of hybridomas which were not
177

secreting antibody. Furthermore they found that the proportion of
non-secreting cells did not change after subsequent cloning suggesting
that non-secreting cells may have been passing through a
non-productive stage of the cell cycle.
Preliminary investigations have been carried out to determine
whether such a phenomenon exists in hybridomas and the findings are
described in the next chapter.
Conclusions.
1. Higher cell and antibody yields from glutamine were achieved in
chemostat culture suggesting that glutamine utilisation
is more efficient
at low concentrations. The lower ammonia yields observed in the
chemostat were also evidence of a more efficient incorporation of
glutamine nitrogen.
2. Glucose uptake rate increased at higher growth rates. The proportion
of glucose converted to lactate was also increased at higher growth
rates suggesting that energy production from aerobic glycolysis
increases in rapidly growing cells.
3. Glutamine uptake rate increased at high growth rates. The amount of
ammonia liberated from each molecule of glutamine declined at high
growth rates suggesting that glutamine nitrogen was being more
efficiently incorporated into biomass at high growth rates.
4. The specific production rate for antibody declined at high growth
rates. This may be evidence that the increased demands on energy and
178

metabolite pools by rapidly growing cells is detrimental to antibody
synthesis. Alternatively it may be evidence that antibody is only
produced during the GL phase of the cell cycle, which is shorter at high
growth rates, and hence represents a smaller proportion of the cell
cycle.
179

6.1 Introduction.
Chapter 6
Synchronised culture.
There is a considerable weight of evidence to suggest that many
metabolic functions, including protein synthesis, are cell cycle-
dependent (Mitchison, 1971; Lloyd et al., 1982; Denhardt et al., 1986).
Many proteins which show cell cycle-dependent expression can be
related to specific functions within the cell cycle, for example
thymidine kinase (Liu et al., 1986) and histone synthesis (Delisle et al.,
1983; Artishevsky et al., 1984) are usually tightly coupled to DNA
synthesis. However the expression of other proteins may not be so
readily linked with the process of cell replication and immunoglobulin
synthesis falls within this category.
Based on the evidence from cell synchronisation experiments
using different synchronisation procedures on different
immunoglobulin secreting cell lines several investigators have
suggested that the bulk of immunoglobulin production Is produced In
the late Gi early S phases of the cell cycle (Byars and Kidson. 1970;
Buell and Fahey. 1969; Takahashi. 1969; Garatun-Tjeldsto et al.. 1976;
Turner et al.. 1985). Other investigators have disputed this finding in
that they found no evidence for such variation with cell cry Je. (Cowan
and Milstein. 1972; Liberti and Baglioni. 1973) although the latter
authors observed that immunoglobulin synthesis was depressed during
mitosis. It was also suggested that the variation was due to the
synchronisation procedure and not a true reflection of protein
expression in normally growing cells. This view was supported by
180

Damiani and collegues (1979) who investigated the production of
immunoglobulin in an adherent myeloma cell line. Using a physical
rather than chemical selection procedure to synchronise the cells they
could find no variation in immunoglobulin expression during the cell
cycle. However Killander et al. (1977) used cytophotometry to
examine the production of IgE in asynchronous cultures of human
myeloma cells and in At least some of their experiments there was an
apparent increase in the accumulation of IgE in Gi with a subsequent
decrease during S phase.
There are therefore several different observations concerning the
relationship between cell cycle and immunoglobulin production but
the observations that the G, phase contributes much of the variability
to the duration of the cell cycle and that cells, including hybridomas
(Schliermann et al., 1987). tend to accumulate in Gi in sub-optimal
growth conditions suggest that higher antibody production rates in G1
could contribute to the variability seen in immunoglobulin production
by hybridomas.
In order to investigate the possibility that immunoglobulin
synthesis or secretion is restricted to particular phases of the
hybridoma cell cycle it is necessary to produce a population of cells
which are predominantly in the same part of the cell cycle at one time.
Cell synchronisation allows this approach in that a population of cells
can be induced to perform the mechanisms of cell division in synchrony.
The methods available for synchronisation of animal cells depends
on the type of cell and a number of methods have been applied.
181

61 .1 Physical selection methods.
1) Mitotic selection. The selection of cells in mitosis is a method which
produces a population of cells with a high degree of synchrony
(Terasima and Tolmach. 1963). The tendency for anchorage dependent
cells to become - rounded and hence more loosely attached to the
substrate during mitosis means that they are more readily shaken
free of the surface and can be separated from the more strongly
attached cells. This technique only produces a small number of cells
and cannot be applied to suspension cells such as hybridomas.
ii) Size selection. In progressing from one mitosis to the next a cell
doubles in size. In fact there Is a gradual Increase in cell volume from
Gi to M (Grazitt et al., 1978) and this has been used to separate cells
by density gradient centrifugation (Mitchell and Tupper, 1977).
iii) Cell sorting by cytofluorimetry. Fluorescence measurements of the
DNA content of cells can be made at rates of 104 cells per second in a
flow cytofluorimeter (Kamentsley et al., 1965) which allows the
frequency distribution analysis of the cell cycle in randomly dividing
cells. Cell subpopulations can thus be selected and the physical sorting
of cells in particular phases can be achieved as the cells leave the flow
chamber. Cells are isolated in tiny droplets and selected cells are
electrostatically charged and deflected into a separate collection vessel
for further analysis (Horan and Wheeless, 1977; Schaap et al., 1979).
182

6.1.2 Growth arrest.
Other synchronisation procedures depend on the accumulation of cells
at a particular point in the cell cycle by the use of inhibitors or
nutrient starvation.
1) Nutrient starvation. Under adverse environmental conditions
mammalian cells have a tendency to accumulate in the Gi phase and this
phenomenon has been used to synchronise cells. For example cells
starved of isoleucine or glutamine (Ley and Tobey, 1970) or serum
(Burk, 1970) arrest in Gi and subsequent addition of the missing
nutrient can induce the cells to divide in synchrony.
11) Inhibition. The use of metabolic inhibitors to specifically block an
event necessary for the progress of the cell through the division cycle
can be used to gather cells at a particular point in the cell cycle.
Release from a reversible inhibitor can result in the synchronous growth
of the cells. The use of colcemid to gather cells at metaphase by the
Inhibition of spindle formation (Stubblefield, 1964). is useful In
gathering greater numbers of anchorage dependent cells for mitotic
selection but is not useful for suspension cells because exposure to
colcemid for longer than 2-3 hours causes irreversible damage to the
cells (Stubblefield, 1968). Inhibitors of DNA synthesis are more useful
for suspension cells. Methotrexate (Rueckert and Mueller, 1960) or
6-fluorodeoxyuridine (Littlefield, 1962) inhibit DNA synthesis by
interfering with the methylation of thymidine. High
concentrations of thymidine (Xeros, 1962). deoxyadenosine or
183

deoxyguanosine (Mueller and Kajiwara, 1966) inhibit DNA synthesis by
feedback effects on other nucleotide precursors. Such methods will
arrest cells at any point in the S-phase and consequently a single block
will only produce a high degree of synchrony in cells where the
S-phase represents only a small fraction of the complete division
cycle. More rapidly growing cells such as hybridomas where the
S-phase occupies a relatively large proportion of the cell cycle
require a double block. The first block arrests cells in S-phase while
cells in other phases complete the cycle and gather at the end of the
Gi phase. The cells are then released from the block until all cells
have traversed the S-phase then blocked a second time such that the
majority of cells collect at the boundary between the Gi and S phases.
On release the cells should resume growth in synchrony.
Thymidine blocking of 321 hybridoma cells has already been
successfully used in this investigation (see Chapter 4). The double
thymidine block method was therefore adapted for the
synchronisation of 321 hybridoma cells.
6.2 Materials and methods.
6.2.1 Synchronisation of the 321 hybridoma.
The synchronisation of the 321 hybridoma was carried out
according to the method of Puck (1964). The method was modified for
the hybridomas.
Cells growing in early to mid exponential phase were arrested
by the addition of thymidine to 2mM. After 30 hours under the
184

thymidine block the cells were washed and resuspended in thymidine
free medium at a density of 2.6x105 cells/ml. After 6 hours incubation
2mM thymidine was added to the culture and the cells were incubated
for a further 18 hours. The cells were resuspended at a density of
2.5x105 /ml in thymidine free medium and the cell growth and
productivity were measured for 2 cell divisions.
6.2.2 Radiolabelling of proteins.
" Glasgows modification of Eagles medium was used for the protein
radiolabelling experiments. This was formulated to contain only
3mg/l of methionine and was obtained from the Institute of Virology,
University of Glasgow.
For labelling experiments an aliquot containing 106 cells was
washed twice with PBS at 37° C and resuspended in 200pl GMEM
containing 35 S-methionine (Amersham) at a concentration of
20liCi/ml. The cells were incubated with the radiolabel for one hour
then lysed by the addition of 200p1 of sample buffer (100mM tris HCl
pH6.8,10% 2-mercaptoethanol, 4% sodium dodecyl sulphate and 20%
glycerol) and each sample was boiled for 6 minutes before storage at
-2011 C.
6.2.3 Radiolabelling of DNA.
Periods of DNA synthesis were assessed by harvesting aliquots
of 5x105 cells and resuspending them in iml RPMI containing
3 H-thymidine (Amersham) at 6pCi/ml. The cells were incubated for 30
185

minutes then collected by centrifugation. The supernatant was
discarded and 0.5 ml distilled water was added to resuspend the cells
which were then lysed by the addition of 0.5m1 SDS-EDTA (0.01M NaCl;
0.01M Tris chloride, pH 7.0; 1% SDS and 0.05M EDTA).
Aliquots of the lysed cell suspension were placed onto filter
paper strips and precipitated with ice cold 10% trichloracetic acid
(TCA). The strips were washed in three changes of 10% TCA. Each
strip was placed in a vial containing 4m1 Aqualuma scintillation fluid
(LKB) and counted in a scintillation counter (LKB Rackbeta).
6.2.4 SDS-Polyacrylamide gel electrophoresis.
SDS-polyacrylamide gel electrophoresis was carried out as
described in section 2.3.4. except that following electrophoresis the gel
was fixed in a solution of 40% acetic acid, 30% methanol in water (v/v)
for one hour then impregnated with En3 Hance (New England Nuclear) for
one hour and finally washed in distilled water for one hour. The gel was
then dried onto filter paper under vacuum.
6.3 Results.
6.3.1 Determination of the S-phase.
The double block method of synchronising animal cells requires
some prior knowledge of the length of S-phase so an experiment was
devised which would give an estimate of the duration of the S-phase.
Cells in the exponential phase of growth were blocked by adding
thymidine to a final concentration of 2mM. After 24 hours the cells
186

were resuspended in thymidine-free medium. At this time and at
subsequent 1 hour intervals, an aliquot of cells was removed and
thymidine was added again to a final concentration of 2mM. After a
further 24 hours incubation in thymidine the cells in each aliquot were
counted. Assuming that there was no delay at the boundary between the
Gi and S phases, the time at which the cell number reaches a plateau
will correspond to the time taken for all cells to escape the S-phase
and hence the duration of the S-phase.
Figure 6.1 shows viable cell counts obtained during the thymidine
block and release experiment. Cells were initially at a concentration
of 3x105 cells/ml before release. After 5 hours a plateau for the cell
concentration was reached indicating that it required approximately
5 hours for all the cells to escape from the S-phase and hence an
approximate duration of 5 hours for the S-phase. Based on this
experiment it was possible to design a protocol for cell synchronisation.
6.3.2 Cell synchronisation.
Figures 6.2 and 6.3 show the results of synchronisation
experiments with the 321 hybridoma. The data shown in figure 6.2 was
the result of a single synchronisation experiment whilst figure 6.3
represents the mean of three experiments. Figure 6.2 is shown
separately as it was the highest degree of synchrony achieved with
the 321 cells.
In each experiment the cells were released from the thymidine
- block at a density of 2.5x105/ml. The cell number remained constant
for 9 hours after which time there was a rapid increase in cell number
187

Figure 6.1. Determination of the duration of S-phase
E
.o
1
0
IC-
C-
a
c
cii
u
o""
o"
0"
0"
0123456789 10
Time from release (hours)
Cells were blocked with 2mM thymidine for an initial period of 24 hours.
The graph shows the effect of of releasing cells from the thymidine block
for increasing periods of time before blocking again with 2mM thymidine.
188

Figure 6.2. Cell growth and antibody production after release from the
double thymidine block.
The figure shows a) viable cell numbers, b) antibody concentration and
c) the specific antibody production rate (APR) calculated from the data
shown in figures 6.2a and 6.2b.
189

x
8
7
Li 3
E
Q
2
140
120
100
80
60
40
20
0
14
12
10
6
a- 6
4
2
Figure 6 .2
a
b
0
0 0 10 20 30 40
Time (hours)
190
ý,

Figure 6.3. Cell growth and antibody production after release from the
double thymidine block.
The figure shows a) viable cell numbers, b) antibody concentration and
c) the specific antibody production rate (APR) calculated from the data
shown in figures 6.3a and 6.3b. Each point represents the mean of three
separate experiments. Error bars are the standard deviation (n=3).
191

E5
8
7
6
Li 3
2
80
70
60
E
öm50
ö 40
0
30
20
10
0
10
8
6
4
2
0
Figure 6.3
0 10 20 30
Time (hours)
192
40

due to a high proportion of the cells dividing simultaneously. The
cells reached a new constant level by which time 70-75% of the cells
had undergone mitosis. The proportion of cells dividing at any one
mitosis never exceeded 75% in these experiments and there was
no significant accumulation of dead cells suggesting that the shortfall
was not due to cell death. The second division occured some 13 hours
later with approximately 70% of the cells undergoing mitosis. There
was considerable loss of synchrony at the second cell division and
this was especially noticeable in figure 6.3a.
It was possible with the Information generated with these
experiments to make estimates for the duration of some of the phases
in the 321 cell cycle. The difference in the time to the first division
and the time between the first and second divisions represents the
duration of the Gi phase assuming that the duration of (S+G2 +M) was
constant in subsequent division cycles. The point at which 50% of the
cells had undergone mitosis was taken as the end of the division
cycle. The approximate duration for Gi from these data was therefore
3 hours and because the duration of the S-phase had been previously
determined as 5 hours the G2 +M phases were regarded as occupying the
remainder of the cell cycle which was 5 hours.
Antibody accumulation in the medium was assayed by the
sandwich ELISA as previously described (Chapter 2, Section 2.2.5).
Figures 6.2b and 6.3b show the accumulation of antibody in the medium
of a synchronised culture of the 321 hybridoma. The rate of
production of antibody was not constant because there were detectable
gradient changes on the antibody curve. Figure 6.2c and 6.3c
Illustrate the variation in antibody production rate during the 321
193

cell cycle. The antibody production rate (APR) was expressed In jig/106
cells/h and was calculated from the cell number and antibody
accumulation curves. Both curves show decreased production rates
coinciding with the M phase and in figure f. 2c there appears to be
a second decrease in production rate which corresponds with the
estimated position of the S-phase. In figure 6.3c the second decrease
In accumulation rate is less apparent which may be the result of the
poorer synchrony achieved in these experiments.
These data however are based only on the accumulation of
antibody in the culture medium and give no indication of the rate of
synthesis of antibody within the cell. To investigate the rate of
synthesis of antibody throughout the cell cycle, the
incorporation of radiolabelled methionine into antibody was
monitored at intervals throughout the cell cycle.
6.3.3 The incorporation ofS-methionine into antibody.
A synchronised culture was initiated at 2.5x105 cells/ml and aliquots
of cells were removed at 90 minute intervals to measure the rate of
incorporation of 35 S-methionine into 321 antibody.
Figure 6.4a shows cell number in a synchronised culture of 321 cells
over a period of 18 hours. Figures 6.4b and 6.4c show respectively, the
incorporation of 3 H-thymidine Into DNA and the incorporation of
30 S-methionine into protein over the same period. The rate of
incorporation of thymidine was initially high but decreased after
approximately 5 hours which is in agreement with previous estimates of
the duration of the S phase. There was a subsequent increase in the
194

Figure 6.4. Incorporation of 3 H-thymidine and 35 S-methionlne býsynchronised
hybridomas.
The figure shows a) viable cell numbers, b) 3 H-thymidine incorporation
into acid precipitable material and c) 35 S-methionine incorporation into
acid precipitable material.
195

u,
X
53
N
G!
Li
°5
x
w
Ln
03
r
4
2
6
cu 4
Li
a- 2
V
0
;n 14
0 12
N
ü 10
Ln
08
CL- 6
, __, 4
Figure 6.4
a
p
0
O
b o
c
024 6' 8 10 12 14 16 18
0
0
0
Time from release (hours)
CPM=
counts per minute
196

ate of thymidine incorporation with a 60% maximum point 13 hours
from the release from the thymidine block.
Plate 6.1 shows an autoradlograph developed from an
SDS-polacrylamide gel of radlolabelled proteins from a synchronised
culture of 321 cells. The heavy chain of the antibody was obscured by
other bands around the same position on the gel and its relative
density could not be readily determined. The light chain however was
a easily distinguishable band with a molecular weight of 25kD. The
relative density of the light chains was determined using a Quantimet
970 image analyser (Cambridge Instruments) programmed to measure
grey levels along an individual track. The variation in intensity
between tracks was compensated for by dividing the intensity of the
light chain peak by the average intensity of the entire track. The
resulting graph (figure 6.5) shows that the incorporation of
35 S-methionine into antibody light chains shows a similar periodicity "
to the accumulation of antibody in the medium. Furthermore because
both intracellular and secreted antibody was measured by this method
it suggests that the observed fluctuations are due to the rate of
antibody synthesis. The relative rates of antibody synthesis and
secretion were not determined in this investigation but could be
measured by pulse-chase radlolabelling experiments.
6.3.4 The effect of PH on the cell cycle.
In order to explain the variation in antibody production rate due
to growth rate as a result of the periodic production of antibody within
the cell cycle it would be necessary to demonstrate that variations in
cell growth rate can be attributed to variations in the duration of
197

Plate 6.1 Autoradlograph of the separated 321 hybrldoma proteins.
The plate shows an autoradlograph of the 321 hybridoma proteins
which have incorporated 35S-methionine during a 60 minute pulse
labelling of cells in a synchronous division cycle. Each lane represents
a separate time interval after release from the thymidine block as shown
by the times above the individual lanes. Molecular weight markers are
14 C-labelled proteins (myosin, 200kD; phosphorylase-b, 92.6kD; bovine
serum albumin, 69kD; ovalbumin, 46kD; carbonic anhydrase, 3OkD;
lysozyme, 14.3kD; Amersham CFA 626).
198

Plate 6.1
Molecular weight
markers (kD)
45-- raw
29
Time from release (hours)
o Ln CD Ln D Ln
rNLn CO
°
Ln 9 c. n o9
om Co
r- v- r-
F "' Ir p'I * '+
199
-. "-_- --
H
-H
-L

Figure 6.5 Changes in the incorporation of 35 S-methionine into
antibody light chains.
a-
a,
2.4
2.2
2.0
1.8
1.6
v 1.4
1.2
1.0
02468 10 12 14 16 18
Time from release (hours)
The figure shows the relative density of the light chain bands observed
on an autoradlograph of a PAGE gel loaded with protein samples from the
synchronised culture described in figure 6.4. The relative density of the
bands was determined by image analysis using the Quantimet image
analyser (Cambridge Instruments).
200

particular phases of the cell cycle.
It has previously been shown that pH had a significant effect on
the growth rate of the 321 hybridoma (Chapter 4, Section 4.2.4). For this
reason the effect of pH on the cell cycle was investigated using a
synchronised culture of 321 cells. Synchronised cells were prepared
as previously described and were resuspended in HEPES buffered RPMI
at initial pH values of 6.6,7.2. and 7.8.
Figure 6.6 shows cell counts taken over 34 hours. It was observed
that the cells underwent the first mitosis after release simultaneously
and reached the same cell density. This implies that the S, G2, and M
phases seem unaffected by a this pH range. However at the second
mitosis cells at pH 7 underwent division some hours earlier than those
cells at pH 7.8 whilst cells at pH 6.6 underwent division some hours
later than cells at pH 7.8. If it is assumed that the S, G2 and M phases
are still constant then the Gi phase appears to be contributing to the
Increase In length of the cell cycle under unfavourable conditions of pH.
6.4 Discussion.
Synchronisation of the 321 hybridoma was achieved using the
double thymidine block method. The cell cycle was approximately 13
hours in length under these conditions and this agreed favourably with
the maximum specific growth rate in batch culture of 0.051 h-1 which is
equivalent to a cell doubling time of 13.6 hours. Analysis of the data
from several experiments showed that the S phase was 5 hours long, the
combined G2 +M phase 5 hours long suggesting that the minimum duration
of the Gi phase was 3 hours.
201

Figure 6.6. The effect of pH on the duration of the cell cycle
Ln
ö
Irl-
x
-4 E
-i a
u2
8
7
6
1
0 0 10 20 30 40
Time from release (hours)
The figure shows viable cell numbers determined for synchronised
hybridoma cultures at pH 6.6 (0). pH 7.2 (0) and pH 7.8 (Q). The cells
were cultured in RPMI 1640 with 10% newborn calf serum, 5mM NaHCO3
and 20mM HEPES. The pH was, adjusted to the above values by the
addition of NaOH or HCI.
202

The data generated from cell synchronisation experiments on
the 321 hybridoma suggested that the rate of production of
immunoglobulin was not constant throughout all phases of the cell
cycle. In particular there was a marked and reproducible drop in the
antibody accumulation rate during mitosis with evidence of a second
decrease in the S phase.
Depression of protein synthesis during mitosis has been described
in many mammalian cells (eg. Prescott and Bender, 1962; Scharff and
Robbins, 1966; Lloyd et al., 1976). This may be due in part to the well
documented stop of RNA synthesis at mitosis (Feindengen et al., 1960;
Terasima and Tolmach, 1963; Schiff, 1965) but some cell lines have
been shown to continue protein synthesis during mitosis (Konrad,
1963). Evidence has been produced to suggest that where a decrease
in protein synthesis occurs it is due to inhibition of the translation
of mRNAs which persist in the cytoplasm during mitosis (Fan and
Penman, 1970).
In this Investigation total protein synthesis was assessed by
the Incorporation of 35 S-methionine Into TCA precipitable protein.
On release from the thymidine blockade there was an Initial increase
in the incorporation of the radiolabel until mitosis at which point
there was a 35% decrease. The rate of 35 S-methlonlne Incorporation did
not recover after mitosis but continued at the lower level. It was
therefore difficult to determine whether there was a decrease in the
rate of protein synthesis during mitosis.
The incorporation of 3a S-methionine into Immunoglobulin
light chains however, showed a decrease during mitosis with a
203

subsequent increase in the following Gi phase. This is in agreement
with the observations of several authors who noted that the
production of antibody was lowest during mitosis (Takahashi et al.
1969; Buell and Fahey, 1969; Lerner and Hodge, 1970; Libertl and
Baglioni, 1973; Turner et al., 1985). Several of these workers also
showed that peak rates of antibody production in late Gi and during S
phase were followed by a gradual decrease during G2 and mitosis: in
both human lymphoblastoid cells (Buell and Fahey, 1969; Takahashi et
al., 1969; Turner et al., 1985) and, mouse myeloma cells (Byars and
Kidson, 1970; Garatun-Tjeldsto et al., 1976).
However, in contrast to the patterns described above, the
antibody accumulation rate In the synchronised hybridoma remained
high during the latter half of the cell cycle. In one experiment two
peaks in the rate of antibody accumulation could be resolved. These
peaks corresponded to the estimated positions of the Gi and G2 phases
with an apparent decrease in the antibody production rate during S
phase. This is consistent with - the patterns of immunoglobulin
production which have been observed in human myelomas by Killander
et al. (1977) and in human lymphoblastold cells by Watanabe et al
(1973). These authors found high production rates in Gi, decreased
production rates during S phase and a later recovery in G2.
Many experiments in this investigation did not show this decrease
in antibody production rate during S-phase although there was an
abrubt increase -corresponding to the G2 phase. However in the
radiolabelling experiment, there was an apparent reduction in the
incorporation of 3° S-methionine into antibody light chains coinciding
with the peak rate of 3H-thymidine incorporation into DNA. This
204

suggests that there was a reduction in antibody production during the
S phase which may have been obscured in some experiments by a
lower degree of cell synchrony. However the reduction in the rate of
antibody synthesis during S phase must be either partial or confined to
a small portion of the S phase. If antibody production stopped completely
for the duration of the S phase (40% of the cell cycle) the change In the
rate of antibody accumulation might be expected to be greater than
observed in this Investigation.
The Increased rate of immunoglobulin synthesis seen In G2 In many
experiments may be a reflection of increases in cell size and protein
content due to cell growth. The dry weight of 'a cell increases
continuously during the cell cycle and pulse labelling
with amino acids shows an increasing rate of incorporation
(Zetterberg and Killander, 1965). The rate of RNA synthesis also
doubles between mitoses (Scharff and Robbins, 1966; Enger and Tobey.
1969; Lloyd et al., 1982). This suggests that the specific rate of
protein synthesis per cell increases and may account for the higher rate
of antibody accumulation during the G2 phase. The incorporation of
35S-methionine did show an initial increase in the rate of uptake on
progression from S to M in the first cycle but this was not repeated in
the second cycle. This may have been due to nutrient depletion causing
a slowing of protein synthesis.
The abrupt increase in antibody accumulation rate which was
frequently observed immediately after the S phase might also be
explained by a gene dosage effect. Several authors have noted that the
rate of RNA synthesis doubles during S phase and this has been ascribed
to a doubling of the gene complement after DNA replication which leads
205

to Increased rates of transcription (Zetterberg and Killander, 1966;
Stambrook and Sisken, 1972). Further investigation would be required
to confirm this hypothesis, such as the measurement of heavy or light
chain immunoglobulin messenger RNAs by suitable cDNA probes to
show whether there is increased transcription of immunoglobulin genes
following DNA duplication. The same approach might equally be
applied to all phases of the cell cycle to investigate whether the
levels of immunoglobulin mRNA reflect the pattern of
immunoglobulin synthesis during the cell cycle.
The validity of data generated from synchronisation
experiments has been questioned by a number of authors
(Mitchison, 1971; Cowan and Milstein, 1973, John et al., 1981; Lloyd,
1986) due to the techniques used to obtain synchronised cells. In
particular methods of Induction synchrony require the maintenance of
cells in restrictive conditions often for long periods of time. It Is
known that many of the inhibitors of DNA synthesis, including
thymidine (Xeros, 1962), allow continued RNA and protein synthesis
and it has been suggested that blocking with inhibitors of DNA
synthesis only synchronises those events which depend on or are
coupled to DNA synthesis (Mueller, 1969) whilst protein and RNA
synthesis remain asynchronous. This suggests that cell growth and DNA
synthesis can be uncoupled from one another as first proposed by
Mitchison (1971) however there is a considerable amount of evidence
which suggests that the intiation of DNA synthesis is dependent
ultimately on an event linked to cell size (Shields et al., 1978, Nurse and
Fantes. 1981).
With due regard to reservations concerning cell
synchronisation methods the data presented In this investigation
206

indicate that antibody production rate is not constant throughout the
cell cycle. However further work is required in order to confirm that the
variations observed are not a result disturbances in cell metabolism due
to the synchronisation method. Methods of synchronising cells which do
not rely on perturbation of cell metabolism (e. g. density gradient
separation) will provide a more accurate impression of antibody
production patterns during the cell cycle. In an attempt to overcome
perturbations introduced by cell synchronisation techniques, Altshuler
et al.. 1986 examined the immunoglobulin content of an asynchronous
population of hybridomas by flow cytometry. They observed two peaks
in the intracellular concentration distribution of immunoglobulin which
might suggest a cell cycle dependence for antibody synthesis. However
they also showed with dual staining techniques for DNA and antibody
content that both antibody peaks appeared to be present throughout
the cell cycle and concluded that there was no simple on/off
dependence of immunoglobulin production on cell cycle. However flow
cytometry only measures the amount of antibody associated with a cell
at a particular point in time and does not provide Information on the
rate of antibody production. Thus their investigation may only have
demonstrated the existence of two populations of hybridoma cells.
It has not been ascertained' in this Investigation whether
variations in the rate of antibody production were due a temporal
regulation of antibody synthesis during the cell cycle or were a
reflection of the changing demands for macromolecular synthesis during
certain phases of the cell cycle. For example many enzymes involved in
DNA synthesis increase in activity during S phase (Kit, 1976; Nakamura
et al., 1984; Eriksson et al., 1984; Denhardt et al., 1986) and histone
synthesis may account for 10% of total protein synthesis during this
207

period (Denhardt et al., 1986). This could account for a decrease in the
rate of immunoglobulin synthesis during S phase.
Variations in the rate of uptake of nutients during the cell cycle
could also account for variations in the rate of antibody synthesis.
Increased rates of amino acid uptake have been described during Gt' in
human lymphocytes (Glassy and Furlong, 1981) and during mitosis in
HeLa cells (Macmillan and Wheatley, 1981). Further studies on the rates
of nutrient uptake, energy balance and metabolite pools during the
hybridoma cell cycle may clarify the relationship between cell growth
and monoclonal antibody production. Studies on the relative rates of
immunoglobulin gene transcription and translation would be necessary
to reveal whether antibody synthesis was regulated at the gene level or
at a later stage.
The other aim of this Investigation was to determine whether
changes in cell growth rate were attributable to changes in the length
of the Gi phase. It was found that conditions of adverse pH which have
been shown to reduce the growth rate of the 321 hybridoma had little
affect on the time taken to traverse the cell cycle when started from a
point after the Gi phase but significantly lengthen the time taken
to complete the cell cycle when the Gi phase Is present. This Implies
that the Gi phase is more sensitive to the physiological environment
which is in accord with the proposed restriction point of Pardee (1974)
who showed that several different blocking conditions (low serum,
amino acid starvation, elevated intracellular AMP and density
dependent Inhibition) arrested cells at the same point In G1.
These findings are consistent with the idea that changes in cell
growth rate are due to variations in the length of the Gi phase and that
208

the proportions of cells in different phases of the cell cycle must
therefore change with growth rate. This has been demonstrated In yeast
cells growing in chemostat culture where the proportion of cells in the
Gi phase was found to increase at low growth rates (Sheper et al.. 1987).
This implies that the rate of synthesis of proteins which are linked to
particular phases of the cell cycle, will also change with growth rate. It
was not possible to make accurate estimates of the magnitude or
duration of the changes in antibody production rate observed during the
cell cycle of the 321 hybridoma due to imperfect synchrony and therefore
any models of antibody production kinetics based on such data would be
tentative. However it does illustrate that environmental parameters
which affect cell growth rate may have profound effects on the
distribution of cells in the cell cycle which in turn may affect the
relative rate of production of any cell cycle associated proteins. The
relatively high rate of antibody production in Gi would be consistent
with the observation that Gi arrested cells or more slowly growing cells
exhibit higher rates of antibody production.
6.5 Conclusions.
The purpose of this investigation was to determine,
1) whether immunoglobulin synthesis was associated with the G1 phase
of the cell cycle as shown in other Immunoglobulin synthesising cell
lines and,
11) if cell cycle regulated immunoglobulin synthesis could account for
the increased rate of immunoglobulin synthesis observed at low growth
209

ates due to changes in the distribution of cells in the various phases
of the cell cycle.
In an attempt to answer these questions the 321 hybridoma was
synchronised by the double thymidine block technique. Using this
technique It has been found that whilst there is evidence of variation
in the rate of immunoglobulin synthesis at different phases of the cell
cycle. immunoglobulin synthesis did not appear to be restricted to the
Gi phase. Antibody production was observed during the Gz phase but
may be produced at a lower rate during the S phase than during the Gi
or Gz phases. There was a reproducible drop in the rate of antibody
production during the M phase which may be due to a general reduction
in the rate of protein synthesis during this phase although this has not
been conclusively demonstrated in this investigation.
The consequences of such variation on monoclonal antibody
production kinetics have not been ascertained unequivocally but it has
been found that conditions which are detrimental to cell growth such as
low or high pH Increase the length of the cell cycle in a manner which
suggests that this was due to an increase in the length of the Gi phase.
This in turn suggests that the cell cycle distribution changes with
growth rate and that the relative rate of production of cell cycle
associated proteins may change accordingly. This is therefore a possible
mechanism which could explain the effects of certain environmental
parameters such as pH. oxygen or serum concentration on antibody
production by hybridomas.
210

7.1 Summary.
Chapter 7
Summary and Recommendations
A survey of the literature has shown that the kinetics of antibody
production by hybridomas varies greatly between different cell lines.
Many hybridomas show continued antibody production during both the
growth and decline phases of a batch culture and are characteristically
high producers (see for example Birch et al., 1987, Reuveny et al..
1986a) while others show declining rates of antibody production during
the growth phase and are characteristically
low producers (Merten et al.,
1987). While many of these differences may be due to the characteristics
of the myeloma and B-cell parent cells and any genetic recombination
following cell fusion (Westerwoudt, 1986), some generalisations can be
made. For example it has been found that restriction of the rate of cell
proliferation by nutrient limitation or by other suboptimal physiological
conditions, often enhances the specific rate of antibody production by
hybridoma cells. This suggests that there is an underlying relationship
between cell growth and antibody production in hybridomas but the
mechanisms of such a relationship have yet to be elucidated. The
purpose of this investigation was therefore to examine the effects of the
culture environment on cell proliferation, cell metabolism and antibody
production in an attempt to determine the relationship between
hybridoma growth and antibody production.
It was considered that the relationship between antibody
production and cell growth could best be evaluated by the use of an on-
line assay to monitor continuously product formation under a variety of
211

experimental conditions. In chapter 3, the development and evaluation
of an automated on-line ELISA system for the determination of antibody
concentrations in bioreactors was described. The assay employed a
packed bed of Sepharose covalently linked to anti-mouse IgG and all
the reaction and washing steps of the ELISA were effected by passing
the reagents through the packed bed. It was demonstrated that such a
system could produce accurate and reproducible data. However the
application of the system to the monitoring of product formation in
bloreactors was limited by the compression of the Sepharose bed after
several assay cycles which reduced the accuracy of the assay. Further
difficulties were experienced in producing a cell-free sample stream for
the assay while maintaining the sterility of the reactor. The
recirculation of cells and medium through a cross-flow filter device was
found to be detrimental to cell growth.
In chapter 4 the growth and antibody production kinetics for the
321 hybridoma were examined. The 321 hybridoma was found to produce
large amounts of antibody with titres of 120-200pg/ml achieved in batch
culture and in common with other highly productive hybridomas, the
production of antibody continued during the decline phase. Furthermore
the specific rate of antibody production was higher during the decline
phase and approximately 60% of the total antibody yield was produced
during this phase. Glutamine was exhausted by the end of the growth
phase suggesting that glutamine may not be necessary for the continued
production of antibody. One possible explanation for the increased
antibody production rate after glutamine depletion was that glutamine
or products of glutamine metabolism may have been Inhibiting or
repressing antibody synthesis. However while lower giutamine
concentrations led to decreased cell numbers illustrating the importance
212

of glutamine for hybridoma growth (figure 4.3), an inhibitory effect of
glutamine on antibody production was not apparent. Indeed, there was
some evidence to suggest that antibody production was depressed at low
glutamine concentrations. The continued production of antibody after
the exhaustion of glutamine in batch cultures might be explained by the
accumulation of amino acid pools during growth (Seaver et al., 1984,
Merten et al.. 1986). These pools may be subsequently utilised for
protein synthesis after depletion of exogenous amino acids such as
glutamine (see for example Stoner and Merchant, 1972).
There are several reports in the literature concerning the effect of
serum on the production of antibody by lymphoblastold cell lines (as
recently reviewed by Glassy et al.. 1988) and it has been observed
frequently that antibody production rates are enhanced in serum-free
or serum-depleted media. This has been interpreted by some to suggest
that serum contains factors which are inhibitory to antibody production.
In this Investigation it was found that the reduction of serum
concentration retarded cell growth (figure 4.6a) and that the specific
antibody production rate was enhanced at low serum concentrations
which was in agreement with previous reports. However when cell
proliferation was arrested by the inhibition of DNA synthesis with
excess thymidine, the trend was reversed and the highest antibody
production rates were observed at the highest serum concentration
(figure 4.13). This suggested that the effect of serum on antibody
production was not due to direct inhibition of antibody synthesis but
was an indirect effect due to the effect of serum on cell growth. Glassy
and colleagues have noted that cell growth rates in serum free media are
generally lower than in serum containing media suggesting that low
growth rates are more favourable for antibody production. The data from
213

this investigation is consistent with such a hypothesis.
The optimal pH for growth of the 321 hybridoma was found to be 7.2
which is within the expected range of 6.9-7.4 reported for hybridomas
(Harbour et al.. 1988). However optimal antibody production rates were
found to be at pH 6.8. a phenomenon also reported by Miller et al.
(1988a) suggesting that conditions optimal for cell growth are not
necessarily optimal for antibody production. This effect has also been
noted for oxygen concentration where the optimal oxygen concentration
for antibody production was suboptimal for cell growth (Mizrahi et a).,
1972; Reuveny et al.. 1986a; Phillips et al., 1987). In this investigation
further evidence was provided which suggests that the effects of these
physiological parameters are indirectly linked to antibody production
through their effects on the cell growth rate. For example, when cell
proliferation was arrested with excess thymidine, the optimal pH for
antibody production was pH 7.2 rather than 6.8 indicating that in the
absence of cell growth the most favourable pH for cell proliferation was
also optimal for antibody production. Furthermore It was shown that
preventing cell proliferation by the inhibition of DNA synthesis led to
an increase in the specific antibody production rate (section 4.3.6).
Recently this has been further Illustrated by the observation that cyclic
nucleotides which Inhibit hybridoma proliferation increase the specific
rate of antibody production (Dalill and Ollis, 1988).
Taken together, these observations Imply that low cell growth
rates are favourable for antibody production and that those parameters
which affect cell growth rate also affect the rate of antibody production.
Several reports in the literature are consistent with this hypothesis. For
example in high cell density perfusion culture systems where cell growth
is restricted by nutrient supply, higher specific antibody production
214

ates have been observed (Van Wezel et al., 1985; Reuveny et al., 1986b)
and higher antibody production rates have been associated with low
growth rates in both glucose-limited and oxygen-limited chemostat
cultures (Birch et al., 1985; Miller et al., 1988).
In chapter 5a glutamine-limited chemostat culture of the 321
hybridoma was studied in an attempt to determine the effects of cell
growth rate on antibody production and cell metabolism. In common with
the reports cited above it was found that antibody production rates were
higher at low growth rates (table 5.3) and antibody yields from )
glutamine were also considerably higher at low growth rates. Three ý
possible explanations for this behaviour were put forward.
1) The metabolic costs of cell growth are high and there is greater
competition for the energy and metabolite pools available for
biosynthesis at high growth rates. These increased demands might be
expected to be reflected by changes in cell metabolism and indeed
studies on the steady state metabolism of the 321 hybridoma in
chemostat culture showed that the fate of glutamine and glucose
changed markedly with growth rate. High growth rates were accompanied
by an increased rate of glucose uptake with an increased conversion of
glucose to lactate. This is in agreement with the observations of Hu
et al., 1987 and Luan et al. (1987b) who noted that there was an
apparent relationship between growth rate and the yield of lactate from
glucose in batch cultures of hybridomas. This increased flux through
glycolysis may be a reflection of an increased demand for glycolytic
intermediates in rapidly growing cells as suggested by Hume et al.
(1977). Alternatively Lazo (1981) has suggested that many cells may
derive much of their energy requirements from aerobic glycolysis and the
215

increased lactate yield at high growth rates may be, evidence of
increased energy requirements especially as the alternative energy
source (glutamine) was limiting in these cultures.
The incorporation of glutamine nitrogen into cell biomass was also
more efficient at high growth rates as illustrated by lower ammonia
yields and was further evidence for an increasing demand for metabolic
Intermediates.
2) Products of cell metabolism are inhibiting or repressing antibody
synthesis possibly due to feedback regulation of biosynthetic pathways.
Whether this is the case for the 321 hybridoma has not been ascertained
as only s limited number of metabolites were monitored at steady state.
However there was no evidence that glutamine was inhibitory to
antibody production from the batch culture experiments (chapter 4). The
metabolic quotient for ammonia did not change at the higher growth
rates (Table 5.3) and was therefore unlikely to be responsible for the
reduced antibody production rate. The increased metabolic quotients for
glucose and lactate at high growth rates suggest that repression of
antibody synthesis by glucose or products of glucose catabolism might
occur as suggested by Glacken et al. (1986). However several workers
have examined the effect of glucose and lactate on antibody production
by hybridomas and have found no evidence of any such repression (Low
and Harbour, 1985; Tharakan and Chau, 1986; Reuveny 'et al.. 1986a;
Miller et al., 1989a).
3) Antibody production is cell cycle related and changes in the cell
population distribution at different growth rates may account for the
changes in antibody production rate. The possibility that antibody
216
S'a

production might be restricted to certain phases of the cell cycle was
investigated in chapter 6. In this chapter it was shown that the rate
of production of antibody by a synchronised culture of the 321
hybridoma varied during the cell cycle. Antibody production did not
appear to be restricted to the G1 phase of the cell cycle as reported for
other lymphoblastold cell lines (Buell and Fahey. 1969; Takahashi et al..
1969; Byars and Kidson. 1970; Garatun-Tjeldsto et a!., 1976; Turner et
a!.. 1985) but continued during the G2 phase. There was a reproducible
decline in the rate of antibody synthesis observed at mitosis (figures 6.2
and 6.3) with some evidence of a further decline during the S phase
although this was variable. It was also shown that pH conditions which
were sub-optimal for cell growth, increased the length of the cell cycle
(figure 6.6) in a manner which was consistent with the Idea that the GI
phase contributes much of the variability to cell cycle times (Pardee,
1978; Scheper et a!., 1988). The observation that the rate of antibody
production may be depressed during the S and M phases when coupled
with the observation that the proportion of cells in these phases is
increased in rapidly growing cells (Schliermann et al., 1987) suggests
that cell cycle related antibody expression cannot be discounted as a
possible mechanism for the growth related effects observed in this and
other studies on immunoglobulin production by hybridomas.
In conclusion it has been found that physiological conditions which
restrict the growth of hybridomas generally enhance the rate of
antibody synthesis. The precise relationship between growth rate and
antibody production rate has not been elucidated but may be a reflection
of the changes in cell metabolism at low growth rates which favour
antibody synthesis. The optimisation of antibody production in large
217

scale bloreactors might therefore be achieved by designing the process
to maintain the cells at the most favourable metabolic state for antibody
production. This might be achieved by the controlled feeding of a
limiting nutient either in a batch or, continuous flow system. The
controlled feeding of nutrients also increases the efficiency of the
process as demonstrated by increased yields of both cells and antibody
in chemostat culture and reduces the accumulation of waste metabolites
such as ammonia and lactate which may be inhibitory.
7.2 Recommendations.
Much of the evidence presented in this invetigation strongly rý
suggests that factors which influence the growth rate of the 321
hybridoma also affect the rate of antibody, production although the
explicit mechanisms behind this relationship have not been elucidated.
Future studies would therefore be centred around the definition of this
relationship and a number of approaches might be considered.
1) Studies on the steady state metabolism of hybridomas in chemostat
culture
It has been hypothesised above that the increased rate of antibody
production at low growth rates is due to a more favourable metabolic
state for antibody production. At steady state the effects of specific
metabolites on cell metabolism and antibody production could be
monitored independently of growth rate. ' Changes in the steady
state concentrations of nutrients such as glucose, glutamine or other
amino acids should affect the relative flux through different metabolic
pathways as suggested by several studies on cell metabolism (Reitzer et
218

al.. 1979). Changes in the metabolic balance could be identified by
monitoring the products of metabolism such as lactate. ammonia,
alanine, pyruvate and ATP and by changes in oxygen uptake. For
example Glacken et al. (1986) have suggested that the relative yields
of ATP from glycolysis or oxidative metabolism can be calculated by the
yield of lactate from glucose and by oxygen uptake rates. Miller and
colleagues have recently reported on the results of step and pulse
changes of glucose, Blutamine, lactate and ammonia in steady state
cultures of hybridomas and have shown that such an approach provides
an insight into the regulation of cell metabolic pathways (Miller et al.,
1988aß 1988b, 1989). The effects of these changes on antibody
production was not assessed by these authors but such an approach
should prove invaluable in determining the effect of the metabolic state
of the cell on antibody production. The effects of possible inhibitory
metabolites such as lactate and ammonia could also be assessed
independently of any effects on cell growth using the chemostat.
2) Studies to identify the level at which antibody production is
regulated.
The possibility of cell cycle regulated expression of immunoglobulin
genes in hybridoma cells has not been ruled out (see Chapter 6).
Furthermore it has been shown that immunoglobulin mRNA may be more
efficiently translated under nutrient starvation (Sonenshein and
Brawerman, 1976) suggesting that post-transcriptional control of
antibody synthesis may occur. Monitoring of the relative rates of gene
transcription and translation should reveal whether immunoglobulin
synthesis is regulated at the gene level or at a later stage by nutrient
or energy limitation. Gene transcription could be monitored by measuring
219

immunoglobulin messenger RNA levels by suitable cDNA probes and
translation by the rate of incorporation of radiolabelled amino acids into
antibody. The possibility of nutrient or catabolite repression of
immunoglobulin genes could also be investigated by monitoring the
relative rates of gene transcription in the presence of high and low
concentrations of metabolites such as glucose or lactate..
3) Studies on the hybridoma cell cycle.
It has not been ascertained whether the variation in antibody
production rate observed in synchronised cultures of the 321 hybridoma
was due to a genuine cell cycle regulation of antibody synthesis or due
to perturbations caused by the synchronisation method. Studies on the
cell cycle of hybridomas by flow cytometry using double labelling
techniques for DNA and antibody content as described by Altshuler et
al., 1987 may provide evidence of cell cycle related immunoglobulin
expression. However such a technique will not provide information on
the relative rates of synthesis or secretion. It would thus be necessary
to study cell populations at particular, phases of the cell cycle with the
minimum perturbation of cell metabolism. This could perhaps best be
achieved by the use of density gradient centrifugation (Mitchell and
Tupper, 1977) or by fluorescence activated cell sorting. Cell populations
thus selected could be analysed using the molecular techniques
described above.
220

References.
Aasen, P. and Listowsky, I. (1980) Iron transport and storage proteins.
Ann. Rev. Biochem. 49,357-393.
Alberts, A. W., Ferguson, K., Hennessy. S., and Vagelos, P. R. (1974)
Regulation of lipid synthesis in cultured animal cells.
J. Biol. Chem. 249,5241-5249.
Altshuler, G. L., Dilworth, R., Sowek, J. and Belfort, G. (1986)
Hybridoma analysis at the cellular level.
Blotechnol. Bioeng. Symp. 17,725-736.
Altshuler, G. L., Dziewulski, D. M., Sowek, J. A. and Belfort, G. (1986)
Continuous hybridoma growth and monoclonal antibody production In
hollow fibre reactors-separators.
Biotechnol. Bioeng. 28,646-658.
Anson, D. S., Austen, D. E. G. and Brownlee, G. G. (1986) Expression of
active human clotting Factor IX from recombinant DNA clones In
mammalian cells.
Nature 315,683-685.
Ardawi, M. S. M. and Newsholme, E. A. (1982a) Glutamine metabolism In
the lymphocytes of the rat.
Biochem J. 212,835-842.
Ardawi, M. S. M. and Newsholme, E. A. (1982b) Maximum activities of
some enzymes of glycolysis, the TCA, and ketone body and glutamine
utilization pathways in lymphocytes of the rat.
Blochem. J. 208, "143-748.
Arnold, J. (1887) Ueber theilunsvoigange an der wanderzellen ihre
progressiven und regressiven metamorphosen.
Arch. Mikrosk. Ant. 30,205-310.
Artishevsky, A., Delegeane, A. M. and Lee, A. S. (1984) Use of a cell cycle
mutant to delineate the critical period for the control of histone mRNA
levels in the mammalian cell cycle.
Mol. Cell. Biol. 4,2364-2369.
Bailey, J . E., Agar, D. W. and Hjortso, M. A. (1982) In Computer
Applications in: Fermentation Technology, Society of Chemical Industry,
London p. 167.
Baker, P., Knoblock, K., Noll, L., Wyatt, D. and Lydersen, B. (1984) A
serum independent medium effective in all aspects of hybridoma
technology and immunological applications.
Develop. biol. Standard. 60,63-72.
Ballou, C. E. (1982) The yeast cell wall and cell surface In: The Molecular
Biology of the yeast Saccharomyces vol 2. Metabolism and gene
expression (Strathern, J., ed. ) Cold Spring Harbour Laboratory. pp. 63-
92.
221

Barnes, D. and Sato, G. (1980) Methods for growth of cultured cells in
serum-free medium.
Anal. Biochem. 102,255-270.
Barseghian, G., Papoian, T., Hwang, D. L., Roitman, A., and Let-Ran, A.
(1984). Ethanolamine mimics insulin effects in vitro.
Biochem. Biophys. Res. Commun. 121,413-421.
Barton, M. E. (1971) Effect of pH on the growth cycle of HeLa cells in
culture without oxygen control.
Biotechnol. Bioeng. 13,471-492.
. Bast, R. C. Jr., Ritz, J., Lipton, J. M., Feeney, M., Sallan, S. E.. Nathan, D.
G. and Schlossman, S. F. (1983) Elimination of leukemic cells from human
bone marrow using monoclonal antibody and complement.
Cancer Res. 43,1389-1394.
Bebbington, C. and Hentschel, C. (1985) The expression of recombinant
DNA products In mammalian cells.
Trends in Biotechnology 3,314-317.
Bell, E., Sher, S., Hull, B., Merrill, C., Rosen, S., Chanson, A., Asselinean,
D., Dubertret, L., Ceulomb, B. C., Lapiere, C., Nusgens, B. and Nevenx, Y.
(1983) The reconstitution of living skin.
J. Investigative Dermatol. 81,25-105.
Berman, P. W., Gregory, T., Crase, D. and Laskey, L. (1985) Protection from
genetic herpes simplex virus type 2 injection by vaccination with cloned
type 1 glycoprotein D.
Science 227,1490-1492.
Berman, P. W. and Lasky, L. A. (1985) Engineering glycoproteins for use
as pharmaceuticals. Trends in Biotechnology 3,51-63.
Birch, J. R. and Hopkins, D. W. (1977) The serine and glycine
requirements of cultured human lymphocyte lines In: Cell culture and Its
Application (Acton, R. T. and Lynn, J. T., eds. ), Academic Press, New
York. pp. 503-511.
Birch, J. R. and Edwards, D. J. (1980) The effect of pH on the
growth and metabolism of a lymphoblastoid cell line.
Develop. biol. Standard. 46,69-63.
Birch, J. R., Boraston, R. and Wood, L. (1985a) Bulk production of
monoclonal antibodies in fermenters.
Trends in Biotechnology 3,162-166.
Birch, J. R., Thompson. P. W. Lambert, K. and Boraston, R. (1986b)
The large scale cultivation of hybridoma cells producing monoclonal
antibodies In: Large-Scale Mammalian Cell Culture (Feder, J. and
Tolbert, W. R., eds. ) Academic Press, pp. 1-18.
Birch, J. R., Thompson, P. W., Boraston, R., Oliver, S. and Lambert, K.
(1987) The large scale production of monoclonal antibodies in airlift
222

fermenters in: Plant and Animal Cells : Process Possibilities (Webb, C.
and Mavituna, F., eds. ) Ellis Horwood Ltd., Chichester, pp. 162-171.
Bishop, J. M. (1983) Cellular oncogenes and retroviruses.
Ann. Rev. Blochem. 52,301-354.
Bodeker, B. G. D., Hubner, G. E., Hewlett, G. and Schlumberger,
H. D. (1987) Production of human monoclonal antibodies from immobilised
cells in the Opticell culture system.
Develop. biol. Standard. 66,473-479.
Boraston, R., Thompson, P. W., Garland, S. and Birch, J. R. (1984) Growth
and oxygen requirements of antibody producing mouse hybridoma cells
in suspension culture. Develop. biol. Standard. 55,103-111.
Borrebaeck, C. A. K. (1986) In vitro immunization for production
of murine and human monoclonal antibodies: present status.
Trends in Biotechnology 4,147-153.
Borst, P. (1962) The pathway of glutamate oxidation by
mitochondria Isolated from different tissues.
Blochim. Blophys. Acta 67,256-269.
Boullanne, G. L., Hozumi, N. and Shulman, M. J. (1984) Production of
functional chimaeric mouse/human antibody.
Nature 312,643-646.
Bradford, M. M. (1976) A rapid sensitive method for the quantitation of
microgram quantities of protein utilising the principle of protein-dye
binding.
Anal. Biochem. 72,248-254.
Bravo, R. and Cells, J. E. (1980) A search for differential poly peptide
synthesis throughout the cell cycle of HeLa cells.
J. Cell Biol. 84,795-802.
Bree, M. A., Dhurjati, P., Geoghegan. R. F. and Robnett, B. (1988)
Kinetic modelling of hybridoma cell growth and Immunoglobulin
production in large scale suspension culture.
Biotechnol. Bioeng. 32,1067-1072.
Brooks, R. F. (1975) The kinetics of serum-induced initiation of
DNA synthesis in BHK 21/c13 cells and the influence of exogeneous
adenosine.
J. Cell. Physiol. 86,369-378.
Brosemer, R. W. and Rutter, W. J. (1961) The effect of oxygen
tension on the growth and metabolism of a mammalian cell.
Exp. Cell Research 25,101-113.
Buchegger, F., Haskell, C. H., Schereyer, M., Scazziga, B. R.,
Randin, S., Carrel, S. and Mach, J. P. (1983) Radiolabelled fragments of
223

monoclonal antibodies against carcinoembryonic antigen for localization
of human colon carcinoma grafted Into nude mice.
J. Exp. Med. 158,413-427.
Buell. D. N. and Fahey, J. L. (1969) Limited periods of gene expression in
Immunoglobulin-synthesizing cells.
Science 164,1624-1525.
Burk, R. R. (1970) One step growth cycle for BHK21/C13 hamster
fibroblasts.
Exp. Cell Res. 63,309-316.
Burnet, F. M. (1957) A modification of Jernes theory of antibody
production using the concept of clonal selection.
Austral. J. Sc!. 20,67.
Burrows, M. T. (1910) The cultivation of tissues of the chick embryo
outside the body.
J. Amer. Med. Ass. 55,2057-2058.
Butler. M. and Spier, R. E. (1984) The effects of glutamine utilisation and
ammonia production on the growth of BHK cells In microcarrier cultures.
J. Blotechnol. 1,187-196.
Byars, N. and Kidson, C. (1970) Programmed synthesis and export of
immunoglobulin by synchronised myeloma cells.
Nature 226,648-650.
Byers, V. S. and Baldwin, R. W. (1988) Therapeutic strategies with
monoclonal antibodies and immunoconjugates.
Immunology 65,3,329-336.
Calabresi, P., McCarthy. K. L., Dexter, D. L., Cummings, F. J. and Rotman,
B. (1981) Monoclonal antibody production in artificial capillary cultures.
Proc. Am. Assoc. Cancer Res. 22,362.
Campbell, A. M. (1984) Laboratory techniques in biochemistry and
molecular biology, vol 13. Monoclonal antibody technology. (Burdon. R.
H. and van Knippenberg. P. H. ). Elsevier, Oxford.
Carrel, A. and Burrows, M. T. (1911) On the physicochemical regulation
of the growth of tissues. The effects of the dilution of the medium on the
growth of the spleen. J. Exp. Med. 13,562-570.
Ceccarini C. and Eagle, H. (1971) Induction and reversal of contact
inhibition of growth by pH modification.
Nature, New Biol. 223,271-273.
Chang, T. H., Steplewski, Z. and Koprowski, H. (1980) Production of
monoclonal antibodies in serum-free medium.
J. Immunol. Methods 39,369-376.
Cleveland. W. L., Wood, I. and Erlanger, B. F. (1983) Routine large-scale
production of monoclonal antibodies In a protein-free culture medium.
J. Immunol. Methods 56,221-234.
224

Cole, S. P. C., Vreeken, E. H. and Roder, J. C. '(1985) Antibody production
by human x human hybridomas in serum free medium.
J. Immunol. Methods 78,271-279.
Conrad, A. H. (1971) Thymidylate synthetase activity In cultured
mammalian cells.
J. Biol. Chem. 246,1318-1323.
Cooper. P. D., Burt, A. M. and Wilson, J. M. (1958) Critical effect of
oxygen tension on rate of growth of animal cells in continuous
suspended culture. Nature 182,1508-1609.
Cowan, N. J. and Milstein, C. (1972) Automatic monitoring of biochemical
parameters in tissue culture. Studies on synchronously growing mouse
myeloma cells.
Biochem. J. 128,445-454.
Crabtree, H. G. (1929) Observations on the carbohydrate metabolism of
tumours.
Biochem. J. 23,536-645.
Crawford, J. and Cohen, H. J. (1985) The essential role of L-glutamine
in lymphocyte differentiation in vitro.
J. Cell. Physiol. 124,275-282.
Crowley, C. W.. Liu, C. C. and Levinson, A. D. (1983) Plasmid-directed
synthesis of hepatitis B surface antigen in monkey cells.
Mol. Cell. Biol. 3,44-55.
Dalili, M. and Ollis, D. F. (1988) The influence of cyclic nucleotides on
hybridoma growth and monoclonal antibody production.
Biotech. Lett. 10,781-786.
Damlani, G. Cosulich, E. and Bargellesi, A. (1979) Synthesis and
secretion of IgG in synchronised mouse myeloma cells.
Exp. Cell Res. 118,295-303.
Darfler, F. J. and Insel, P. A. (1982) Serum-free culture of resting,
PHA-stimulated and transformed lymphoid cell lines including
hybridomas.
Exp. Cell Res. 138,287-295.
Davidson, R. L., O'Malley, K. A. and Wheeler, T. B. (1976) Polyethylene
glycol-induced mammalian cell hybridisation: Effect of polyethylene
glycol molecular weight and concentration.
Somatic Cell Genet. 2,271-279.
De'Lisle, A. J. Graves, R. A., Marzluff, W. F. and Johnson, L. F. (1983)
Regulation of histone mRNA production and stability In
serum-stimulated mouse 3T6 fibroblasts.
Mol. Cell Biol. 3,1920-1929.
225

Denhardt, D. T., Edwards, D. R. and Parfett, C. L. J. (1986) Gene
expression during the mammalian cell cycle.
Blochim. Blophys. Acta 865,83-125.
Diderholm, H., Stenvist, B., Ponten, J. and Wesslen, T. (1965)
Transformation of bovine cells In vitro after inoculation of Simian virus
40 or its nucleic acid.
Exp. Cell Res. 37,452-459.
Dodge, T. C. and Hu, W. -S. (1986) Growth of hybridoma cells under
different agitation conditions.
Biotech. Lett. 8,683-686.
Dodge, T. C. Ji, G. -Y., and Hu, W. -S. (1987) Loss of viability in
hybridoma culture: a kinetic study.
Enzyme Microb. Technol. 9,607-610.
Donnelly, M. and Scheffler, I. E. (1976) Energy metabolism in
respiration-deficient and wild-type chinese hamster fibroblasts in
culture.
J. Cell. Physiol. 89,39-62.
Dube, S., Fisher, J. W. and Powell, J. S. (1988) Glycosylation at specific
sites of erythropoietin is essential for biosynthesis, secretion and
biological function.
J. Biol. Chem. 263,17616-17521.
Eagle, H. (1955b) The specific amino acid requirements of a mammalian
cell (strain L) in tissue culture.
J. Biol. Chem. 214,839-852.
Eagle, H. (1955a) Nutrition needs of mammalian cells in tissue culture.
Science 122,501-504.
Eagle, H. (1959) Amino acid metabolism In mammalian cell cultures.
Science 130,432-437.
Eagle, H. (1971) Buffer combinations for mammalian cell culture.
Science 174,500-503.
Eagle, H. (1973) Effect of environmental pH on the growth of normal and
malignant cells.
J. Cell. Physiol. 82,1-8.
Eagle, H., Barban, S., Levy, M. and Schulze, H. O. (1958) The utilization
of carbohydrates by human cell cultures.
J. Biol. Chem. 233,551-558.
Eagle, H., Oyama, V. I., Levy, M., Horton, C. L. and Fleischman, R. (1956)
The growth response of mammalian cells in tissue culture to L-glutamine
and L-glutamic acid. J. Biol. Chem. 218,607-616.
226

Eagle, H. and Piez, K. (1962) The population-dependent requirement by
cultured mammalian cells for metabolites which they can synthesize.
J. Experimental Medicine 116,29-43.
Eaton, M. D. and Scala, A. R. (1961) Inhibitory effect of glutamine and
ammonia on replication of Influenza virus in ascites tumor cells.
Virology 13,300-307.
Eigenbrodt, E., Fister, P. and Reinacher, M. (1985) New perspectives on
carbohydrate metabolism in tumor cells in: Regulation of Carbohydrate
Metabolism Vol. II (Beitner, R. Ed. ) CRC, Boca Raton, FL pp. 141-179.
Emery, A. N., Lavery, M., Williams, B. and Handa, A. (1987) Large scale
hybridoma culture In: Plant and Animal Cells : Process Possibilities
(Webb, C. and Mavituna, F., eds. ) Ellis Horwood Ltd., Chichester. pp.
137-148.
Enders, J. R., Weler, T. H. and Robbins, F. C. (1949) Cultivation of the
Lansing strain of poliomyelitis virus in cultures of various human
embryonic tissues.
Science 109,85-86.
Enger, M. D. and Tobey, R. A. (1969) RNA synthesis in Chinese hamster
cells. II. Increase in the rate of RNA synthesis during G1.
J. Cell Biol. 42,308-315.
Engvall, E. and Perlmann, P. (1971) Enzyme-linked immunosorbent assay
(ELISA) : quantitative assay of immunoglobulin G.
Immunochemistry 8,871-874.
Epstein, N and Epstein, M. (1986) The hybridoma technology : I.
Production of monoclonal antibodies in: Advances in Biotechnological
Processes vol. 6 (Mizrahi, A., ed. ) Alan R. Liss Inc. pp. 179-218.
Epstein, N., Kobiler, D. and Epstein, M. (1986) The hybridoma
technology : II. Applications of hybrid cell products : Monoclonal
antibodies and lymphokines In: Advances in Biotechnological Processes
vol. 6 (Mlzrahi, A., ed. ) Alan R Liss Inc. pp. 219-251.
Eriksson, S., Graslund. A., Skog, S., Thelander, L. and Tribukait. B.
(1984) Cell cycle dependent regulation of mammalian ribonucleotide
reductase.
J. Biol. Chem. 259,11695-11700.
Fan, H. and Penman, S. (1970) Regulation of protein synthesis in
mammalian cells. II. Inhibition of protein synthesis at the level of
initiation during mitosis.
J. Mol. Biol. 50,655-670.
Fantes, P. A. and Nurse, P. (1981) Division timing : controls, models and
mechanisms in: The Cell Cycle (John, P. C. L. ed. ) Cambridge University
Press, Cambridge. pp. 11-34.
227

Fazekas de St. Groth, S. and Scheidegger, D. (1980) Production of
monoclonal antibodies : Strategies and tactics.
J. Immunol. Meth. 35,1-21.
Fazekas de St. Groth, S. (1983) Automated production of monoclonal
antibodies in a cytostat.
J. Immunol. Meth. 57,121-136.
Feinendegen, L. E., Bond, V. P., Shreeve, W. W. and Painter, R. B. (1960) RNA
and DNA metabolism in human tissue culture cells studied with tritiated
cytidine.
Exp. Cell Res. 19,443-469.
Feteris, W. A. (1965) A serum glucose method without protein
precipitation.
Am. J. Med. Technol. 31,17-24.
Fischer, A (1948) Amino acid metabolism of tissue cells in vitro.
Biochem. J. 43,491-497.
Fischer, A. et al. (1948) Growth of animal tissue cells in artificial media.
Proc. Soc. Exp. Biol. Med. 67,40-46.
Fleishaker, R. J. and Sinskey, A. J. (1981) Oxygen demand and supply in
cell culture.
Eur. J. Appl. Microblol. Biotech. 12,193-197.
Frame K. K. and Hu, W. S. (1985) Oxygen uptake of mammalian cells in
microcarrier culture response to changes in glucose concentration.
Blotechnol. Lett. 7,147-161.
Fung, Y. K., Crittenden, L. B., Farley, A. A. and Kung, H. J. (1983) Tumor
induction by direct injection of clonal v. src DNA.
Proc. Natl. Acad. Sc!. USA 80,353-357.
Garatun-Tjeldsto, 0., Pryme, I. F., Weltman, J. K. and Dowben, R. M. (1976)
Synthesis and secretion of light-chain Immunoglobulin In two successive
cycles of synchronised plasmacytoma cells.
J. Cell Biol. 68,232-239.
Gazitt, Y. (1979) Early decrease of 2-deoxyglucose and alpha-amino
Isobutyric acid transport are among the first events in differentiating
synchronised murine erythroleukemia cells.
J. Cell Physiol. 99,407-416.
Gerhard, W., Yewdell, J., Frankel, M. E. and Webster, R. (1981) Antigenic
structure of influenza virus haemagglutinin defined by hybridoma
antibodies.
Nature, 290,713-717.
Geyer, R. P. and Chang, R. S. (1958) Bicarbonate as an essential for
human cells in vitro. Arch. Biochem. Blophys. 73,500-506.
228

Gilliland, D. G., Steplewski, Z., Collier, R. J., Mitchell, K. F., Chang, T. H.
and Koprowski, H. (1980) Antibody-directed cytoxic agents : use of
monoclonal antibody to direct the action of toxin A chains to colorectal
carcinoma cells.
Biochemistry 77,4539-4543.
Glacken, M., Fleischaker, R. and Sinskey, A. (1983) Mammalian cell
culture : Engineering principles and scale-up.
Trends in Biotechnology 1,102-108.
Glacken, M. W., Fleischaker, R. J. and Sinskey, A. J. (1986) Reduction of
waste product excretion via nutrient control : Possible strategies for
maximising product and cell yields on serum in cultures of mammalian
cells.
Biotechnol. Bioeng. 28,1376-1389.
Glacken, M. W. (1988) Catabolic control of mammalian cell culture.
Blo/Technology 6,1041-1050.
Glassy, M. C. (1987) Immortalization of human lymphocytes from a tumor-
involved lymph node.
Cancer Res. 47,5181-5188.
Glassy, M. C. and Furlong, C. E. (1981) Neutral amino acid transport
during the cell cycle of cultured human lymphocytes.
J. Cell Physiol. 107,69-74.
Glassy, M. C., Peters, R. E. and Mikhalev, A. (1987) Growth of human-
human hybridomas in serum-free medium enhances antibody secretion.
In Vitro, 23,745-749.
Glassy, M. C., Tharakan, J. P. and Chau, P. C. (1988) Serum-free media In
hybridoma culture and monoclonal antibody production.
Biotechnol. Bioeng. 32,1015-1028.
Glazer, R. I., Vogel, C. I.., Potel, I. R. and Anthony, P. P. (1974) Glutamate
dehydrogenase activity related to histopathological grade of
hepatocellular carcinoma in man.
Cancer Res. 34,2975-2986.
Goding, J. W. (1983) Monoclonal antibodies : Principles and practice.
Academic Press. New York.
Goldstein, G., Schindler, J., Tsai, H., Cosimi, A. B., Russell, P., Norman, D.
et al. (1985) A randomised clinical trial of OKT3 monoclonal antibody for
acute rejection of cadaveric renal transplants.
New Engl. J. Med. 313,337.
Graff. S.. Moser. H., Kastner, 0., Graff, A. M. and Tannenbaum, M. (1965)
The significance of glycolysis.
J. Natl. Cancer Inst. 34,511-519.
Green. H., Kehinde, 0. and Thomas, J. (1979) Growth of cultured human
epidermal cells into multiple epithelia suitable for grafting.
Proc. Natl. Acad. Sci. 76,5665-5669.
229

Griffiths, J. B. and Pirt, S. J. (1967) The uptake of amino acids by mouse
cells (strain LS) during growth in batch culture and chemostat culture:
the influence of cell growth rate.
Proc. Roy. Soc. B 168,421-438.
Griffiths, J. B. (1986) Can cell culture medium costs be reduced?
Strategies and possibilities.
Trends in Biotechnology 4,268-272.
Guilbert, L. J. and Iscove, N. N. (1976) Partial replacement of serum by
selenite, transferrin, albumin and lecithin in haemopoietic cell cultures.
Nature 263,594-695.
Ham, R. G. and McKeehan, W. L. (1979) Media and growth requirements in:
Methods in Enzymology vol 58 (Jakoby, W. B. and Pastan, I. H., eds)
Academic Press, New York pp. 44-93.
Handa, A., Emery, A. N. and Spier, R. E. (1987) On the evaluation of
gas-liquid Interfacial effects on hybridoma viability In bubble column
bioreactors.
Develop. biol. Standard. 66,241-253.
Harris, J. L. and Spier, R. E. (1985) Physical and chemical parameters:
Measurement and control in: Animal Cell Biotechnology (Spier, R. E. and
Griffiths, J. B., eds) Academic Press, London pp. 283-319.
Harrison, R. G. (1907) Observations on the developing living nerve fibre.
Proc. Soc. Exp. Biol. Med. 4,140-143.
Hendry, R. M. and Herrmann, J. E. (1980) Immobilization of antibodies on
nylon for use in enzyme-linked immunoassay.
J. Immunol. Methods 35,285-296.
Himmelfarb, P. Thayer, P. S. and Martin, H. E. (1969) Spin filter culture:
The propagation of mammalian cells in suspension.
Science 164,555-567.
Himmler, G., Palfl. G., Ruker, F.. Katinger, H. and Scheirer, W. (1984) A
laboratory fermenter for agarose Immobilized hybridomas to produce
monoclonal antibodies.
Develop. biol Standard. 60,291-296.
Holley, R. W., Armour, R. and Baldwin, J. H. (1978) Density dependent
regulation of growth of BSC-1 cells in cell culture: Control of growth by
low molecular weight nutrients.
Proc. Natl. Acad. Sci. USA 75,339-341.
Holmes, B. E. and Watchorn, E. (1927) Studies in the metabolism of tissues
growing in vitro. I. Ammonia and urea production by kidney.
Biochem. J. 21.327-334.
Hopkinson, J. (1985) Hollow fibre cell culture systems for economical
cell-product manufacturing.
Bio/Technology 3,226-230.
230

Horan, P. K. and Wheeless, L. L. Jr. (1977) Quantitative single cell
analysis and sorting. Science 198,149-157.
Howard, A. and Pelc, S. R. (1953) Synthesis of desoxyribonucleic acid in
normal and irradited cells and its relation to chromosome breakage.
Heredity Suppl. 6,261-273.
Hsuing, H. M., Mayne, N. G. and Becker, G. W. (1986) High level
expression, efficiency, secretion and folding of human growth hormone
in Escherlchia coll. Blo/Technology 4,991-995.
Hu, W. -S. and Dodge, T. C. (1985) Cultivation of mammalian cells in
bioreactors.
Biotechnol. Progress 1,209-215.
Hu, W. -S., Dodge, T. C., Frame, K. K. and Himes, V. B. (1987)
Effect of glucose on the cultivation of mammalian cells.
Develop. biol. Standard. 66,279-290.
Hume, D. A., Radik, J. L. Ferber, E. and Weidemann, M. J. (1978) Aerobic
glycolysis and lymphocyte transformation.
Biochem. J. 174,703-709.
Hurrell, J. G. R. (1982) Monoclonal hybridoma antibodies: Techniques and
applications. CRC Press, Boca Raton, Florida.
Hyvarinen, A. and Nikkila, E. (1962) Specific determination of blood
glucose with o-toluidine.
Clin. Chem. Acta 7,140-149.
Imamura, T., Crespi, C. L., Thilly, W. G. and Brunnengraber, H. (1982)
Fructose as a carbohydrate source yields stable pH and redox parameters
In microcarrier cell culture.
Anal. Biochem. 124,353-358.
Iscove, N. N. and Melchers, F. (1978) Complete replacement of serum by
albumin, transferrin and soybean lipids in cultures of lipopolysaccharide
reactive cultures of B-lymphocytes.
J. Exp. Med. 147,923-933.
Ismail, A. A. A., West, P. M. Goldie, D. J. and Poole, A. C. (1981) An
improved fully automated continuous flow system for immunoassays.
Ann. Clin. Biochem. 18,287-294.
Israel, M., Chan, H. W., Hourihan, S. L., Rowe, W. P. and Martin, M. A. (1979)
Biological activity of polyoma viral DNA in mice and hamsters.
J. Virol. 29,990-996.
Ito, M. and McLimans, W. F. (1981) Ammonia inhibition of interferon
synthesis.
Cell Biol. Int. Rep. 5,661-666.
231

Jacobs, R. A. and Majerus, P. W. (1973) The regulation of fatty acid
synthesis in human skin fibroblasts. Inhibition of fatty acid synthesis
by free fatty acids. J. Biol. Chem. 248,8392-8401.
Jacobs, K., Shoemaker, C., Rudersdorf, R., Neill, S. D., Kaufman, R. J.,
Mufson, A., Seehra, J., Jones, S. S., Hewick, R., Fritsch, E. F., Kawakita,
M., Shimizu, T. and Hiyake, T. (1985) Isolation and cloning of genomic
and cDNA clones of human erythropoietin.
Nature 313,806-809.
Jaffres, G. J., Fuller, T. C., Cosimi, A. B., Russell, P. S., Winn, H. J. and
Colvin, R. B. (1986) Monoclonal antibody therapy. Anti-idiotypic and
non-anti-idlotypic antibodies to OKT3 arising despite intense
immunosupression.
Transplantation 41,578-582.
James. K. and Bell, G. T. (1987) Human monoclonal antibody production.
Current status and future prospects.
J. Immunol. Methods 100,5-40.
James, R. and Bradshaw, R. A. (1984) Polypeptide growth factors.
Ann. Rev. Biochem. 53,259-292.
Janatova, J. and Gobel, R. J. (1984) Rapid optimisation of
immunoadsorbent characteristics.
Biochem. J. 221.113-120.
Jarvis and Grdina, T. (1983)
microencapsulated living cells.
Biotechniques
Production of biologicals from
1,22-27.
Jensen. E. M. and Liu, 0. C. (1961) Studies of Inhibitory effects of
ammonium ions in several virus-tissue culture systems.
Proc. Soc. Exp. Biol. Med. 107,834-838.
John, P. C. L., Lambe, C. A., McGookin, R. and Orr, B. (1981) Accumulation
of protein and messenger RNA molecules in the cell cycle in: The Cell
Cycle (John, P. C. L., ed. ) Cambridge University Press, Cambridge. pp.
185-222.
Johnson, I. S. (1983) Human Insulin from recombinant DNA technology.
Science 219,632-637.
Johnson, L. F., Fuhrman, C. L. and Wiedman, L. M. (1978) Regulation of
dihydrofolate reductase gene expression in mouse fibroblasts during
transition from resting to growing state.
J. Cell. Physiol. 97,397-406.
Kadourl, A. and Bohak, Z. (1985) Production of plasminogen activator by
cells in culture in: Advances in Biotechnological Processes, Vol 5
(Mizrahl, A. and van Wezel, A. L., eds) Alan R Liss, New York
pp. 275-292.
232

Kamentsky, L. A., Melamed, M. R. and Derman, H. (1965)
Spectrophotometer: new Instrument for ultrarapid cell analysis.
Science 160,630-631.
Karkare, S. B., Phillips, P. G., Burke, D. H. and Dean, R. C. (1985)
Continuous productionof monoclonal antibodies by chemostatic and
immobilized hybridoma culture in: Large-Scale Mammalian Cell Culture
(Feder, J. and Tolbert, W. R., eds) Academic Press, New York. pp 128-150.
Katinger, H. W. D. and Bliem, R. (1983) Production of enzymes and
hormones by mammalian cell culture in: Advances in Biotechnological
Processes, vol. 5 (Mizrahi, A. and van Wezel, A. L., eds) Alan R Liss, New
York. pp. 61-95.
Katinger, H. and Scheirer, W. (1985) Mass cultivation of animal cells in:
Animal Cell Biotechnology. vol 1 (Spier, R. E. and Griffiths, J. B., eds)
Academic Press, London pp. 167-193.
Katinger, H. (1987) Principles of animal cell fermentation.
Develop. biol. Standard. 66,195-209.
Kaufmann, S. H. E., Eichmann, K., Muller, I. and Wrazel, L. J. (1985)
J. Immunol. 134,4123-4127.
Kaufman, R. J., Wasley, L. C., Spiliates, A. J., Gossels, S. D., Latt, S. A.,
Larsen, G. R. and Kay, R. M. (1985) Co-amplification and co-expression
of human tissue-type plasminogen activator and murine dihydrofolate
reductase sequences in Chinese hamster ovary cells.
Mol. Cell. -Biol. 5,1750-1769.
Kawamoto, T., Sato, J. D., McClure, D. B. and Sato, G. H. (1986)
Serum-free medium for the growth of NS- 1 mouse myeloma cells and the
isolation of NS-1 hybridomas In: Methods in Enzymology, vol 121
(Langone, J. J. and van Vunakis, H., eds) Academic Press, London
pp. 266-277.
Kenealy, W. R., Gray, J. E., Ivanoff. L. A., Tribe, D. E., Reed, D. L., Korant,
B. D. and Petteway, S. R. (1987) Solubility of proteins overexpressed in
Escherichla coll.
Developments in Industrial Microbiology 28,45-52.
Kernan, N. A., Byers, V. S., Scannon, P. J., Mischak, R. P., Brockstein, J.,
Dupont, B. and O'Reilly, R J. (1988) Treatment of steroid resistant
graft-vs-host disease by in vivo administration of an anti-T-cell ricin
A chain Immunotoxin.
J. Amer. Med. Assoc. 259,3154-3157.
Kilburn, D. G. and Webb, F. C. (1968) The cultivation of animal cells at
controlled dissolved oxygen partial pressure.
Biotechnol. Bioeng. 10,801-814.
Killander, D., Nilsson, K., Lundin, L. and Fabriclus, H. -A. (1977)
Cytoplasmic levels of IgE in human myeloma cells in different phases of
the cell cycle.
Eur. J. Immunol. 7,786-791.
233

Kim, W., Tsao, D., Itzkowitz, S. H. and Kim, Y. S. (1985) Production and
clinical utility of carcinoembryonic antigen In: Advances In
Biotechnological Processes, vol 4 (Mizrahi, A., and van Wezel, A. L., eds. )
Alan R Liss, New York pp. 151-181.
King, M. E. and Spector, A. A. (1981) Lipid metabolism In cultured cells
In: The Growth Requirements of Vertebrate Cells in vitro (Waymouth, C.,
Ham, R. G. and Chapple, P. J., eds. ) Cambridge University Press, London
pp. 293-312.
Kingsman, S. M.. Kingsman, A. J. and Mellor, J. (1987) The production of
mammalian proteins In Saccharomyces cerevisiae.
Trends In Biotechnology 5,53-67.
Kit, S. (1976) Thymidine kinase, DNA synthesis and cancer.
Mol. Cell. Biochem. 11,161-182.
Kitos, P. A., Sinclair, R. and Waymouth, C. (1962) Glutamine metabolism
by animal cells growing in a synthetic medium.
Exp. Cell Res. 27,307-316.
Klein, F., Ricketts, R., Pickle, D. and Flickinger, M. C. (1983)
Interleukin-2 and interleukin-3: Suspension cultures of constitutive
producer lines in: Advances in Biotechnological Processes, vol 2
(Mlzrahi, A. and van Wezel, A. L., eds. ) Alan R Liss, New York pp.
111-134.
Klement, G., Scheirer, W. and Katinger, H. W. D. (1987) Construction of a
large scale membrane reactor system with different compartments for
cells, medium and product.
Develop. biol. Standard. 66,221-226.
Kluft, C., van Wezel, A. L., van der Velden, C. A. M., Emels, J. J., Verheilen,
J. H. and Wljngaards, G. (1983) Large-scale production of extrinsic
(tissue-type) plasminogen activator from human melanoma cells in:
Advances in Biotechnological Processes Vol. 2. (Mizrahl, A. and van
Wezel, A. L. eds. ) Alan R. Liss, New York. pp 97-110.
Knazek, R. A., Gullino, P. M., Kohler, P. 0. and Dedrick, R. L. (1972) Cell
culture on artificial capillaries: an approach to tissue growth in vitro:
Science 178,65-67.
Kohler, G. and Milstein, C. (1975) Continuous cultures of fused cells
secreting antibody of predefined specificity.
Nature 256,495-497.
Konrad, C. G. (1963) Protein synthesis and RNA synthesis during mitosis
in animal cells.
J. Cell Biol. 19,267-277.
Kovacevic, Z. (1971) The pathway of glutamine and glutamate oxidation
in isolated mitochondria from mammalian cells.
Biochem. J. 125,757-763.
234

Kovacevic, Z. and Morris, H. P. (1972) The role of glutamine in the
oxidative metabolism of malignant cells.
Cancer Res. 32,326-333.
Kovar, J. and Franek, F. (1985) Hybridoma cultivation in defined
serum-free media: Growth supporting substances I. Transferrin.
Folia biologica (Praha) 31,167-175.
Kovar, J. (1986) Hybridoma cultivation in defined serum-free medium:
Growth supporting substances II. Insulin and other hormones and growth
factors.
Folia blologica (Praha) 32.304-310.
Kovar, J. (1987) Hybridoma cultivation in defined serum-free medium:
Growth supporting substances III. Ethanolamine, vitamins and other low
molecular weight nutrients.
Folia biologica (Praha) 33,369-376.
Kris, R. M., Libermann, T. A., Avivi, A. and Schlesinger, J. (1985) Growth
factors, growth factor receptors and oncogenes.
Bio/Technology 3,135-140.
Krolick, K. A., Uhr, J. W. and Vitetta, E. S. (1982) Selective killing of
leukaemia cells by antibody-toxin conjugates: Implications for
autologous bone marrow transplantation.
Nature 295,604-605.
Kromer, E. and Katinger, H. W. D. (1982) Effects of glucose-feeding on
glucose and oxygen requirement, energy charge and interferon
production of human lymphoid cells.
Develop. biol. Standard. 50,349-354.
Kruse, P. F. Jr., Miedema, E. and Carter, H. C. (1967) Amino acid
utilizations and protein synthesis at various proliferation rates,
population densities and protein contents of perfused animal cells and
tissue cultures.
Biochemistry 6,949-955.
Ku, K., Kuo, M. J., Delente, J., Wildi, B. S. and Feder, J. (1981)
Development of a hollow fibre system for large-scale culture of
mammalian cells.
Blotechnol. Bioeng. 23,79-95.
Kuhlmann, W. (1987) Optimization of a membrane oxygenation system for
cell culture In stirred tank reactors.
Develop. biol. Standard. 66,263-268.
Kvamme, E. and Svenneby, G. (1961) The effect of glucose on glutamine
utilization by Ehrlich ascites tumor cells.
Cancer Res. 21,92-98.
Laemmli, U. K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4.
Nature 227,680-685.
235

Lambe, C. A. and Walker, A. G. (1987) Reactor requirements for animal
cells in: Plant and Animal Cells : Process Possibilities (Webb, C. and
Mavituna, F., eds) Ellis Horwood, Chichester. pp. 116-124.
Lavery, M., Kearns, M. J., Price, D. G., Emery, A. N., Jefferis, R. and
Nienow, A. W. (1985) Physical conditions during batch culture of
hybridomas in laboratory scale stirred tank reactors.
Develop. biol. Standard. 60,199-206.
Lavietes, B. B., Regan, D. H. and Demopoulos, H. B. (1974) Glutamate
oxidation of 6C3HED lymphoma: Effects of L-asparaginase on sensitive
and resistant lines.
Proc. Natl. Acad. Sc!. USA 71,3993-3997.
Lazar, A. (1983) Interferon production by human lymphoblastold cells In:
Advances In Biotechnological Processes, vol 2 (Mlzrahl. A. and van
Wezel, A. L. eds. ) Alan R Liss, New York pp. 179-208.
Lazar, A., Reuveny, S., Geva, J.. Marcus, D., Silberstein, L., Ariel, N.,
Epstein, N., Altbaum, Z., Sinai, J. and Mizrahi, A. (1987) Production of
carcinoembryonic antigen from a human colon adenocarcinoma cell line.
1. Large scale cultivation of carcinoembryonic antigen-producing cells
on cylindric cellulosic-based microcarriers.
Develop. Biol. Standard 66,423-428.
Lazo, P. A. (1981) Amino acids and glucose utilization
metabolic pathways in ascites-tumour cells.
Eur. J. Biochem. 117,19-25.
by different
Lehmann, J.. Piehl, G. W. and Schulz, R. (1987) Bubble-free cell culture
aeration with porous moving membranes.
Develop. biol. Standard. 66,227-240.
Lerner, R. A. and Hodge, L. D. (1970) Gene expression In synchronised
lymphocytes: Studies on the control of synthesis of immunoglobulin
polypeptides.
J. Cell. Physiol. 77,265-276.
Levintow, L., Eagle, H. and Piez, K. A. (1957) The role of glutamine in
protein biosynthesis in tissue culture.
J. Biol Chem. 227,929-941.
Levintow, L. and Eagle, H. (1961) Biochemistry of cultured mammalian
cells.
Ann. Rev. Blochem. 30,605-640.
Levy, R. and Miller, R. A. (1983) Tumor therapy with monoclonal
antibodies.
Fed. Proc. 42,2650-2656.
Lewis, M. R. and Lewis, W. H. (1912) The cultivation of chick tissues in
media of known composition.
Anat. Rec. 6.207-211.
236

Lewis, M. R. (1922) Importance of dextrose in the medium for tissues
cultures.
J. Exp. Med. 53,317-322.
Ley. K. D. and Tobey, R. A. (1970) Regulation of DNA synthesis in chinese
hamster cells. II. Induction of DNA synthesis and cell division by
isoleucine and glutamine in G1-arrested cells In suspension
culture.
J. Cell Biol. 47,453-459.
Liberti, P. and Baglioni, C. (1973) Synthesis of immunoglobulin and
nuclear protein in synchronised mouse myeloma cells.
J. Cell. Physiol. 82,113-120.
Lim, F. and Sun, A. (1980) Microencapsulated Islets as bloartificial
endocrine pancreas. Science 210,908-910.
Littlefield, J. W. (1964) Selection of hybrids from matings of fibroblasts
in vitro and their presumed recombinants.
Science 145,709-710.
Littlefield, J. W. (1966) The periodic synthesis of thymidine kinase in
mouse fibroblasts.
Biochim. Biophys. Acta 114,398-403.
Lloyd, D. (1986) Biochemistry of the cell cycle.
Biochem. J. 242.313-321.
Lloyd. D., Poole, R. K. and Edwards, S. W. (1982) The Cell Division Cycle:
Temporal Organization and Control of Cellular Growth and Reproduction.
Academic Press, London.
Lord, J. M. Roberts, L. M., Thorpe, P. E. and Vitetta, E. S. (1985)
Immunotoxins.
Trends In Biotechnology 3,176-179.
Low, K. and Harbour. C. (1985a) Growth kinetics of hybridoma cells: 1.
The effects of varying foetal calf serum levels.
Develop. biol. Standard. 60,17-24.
Low, K. and Harbour. C. (1985b) Growth kinetics of hybridoma cells: 2.
The effects of varying energy source concentrations.
Develop. biol. Standard. 60,73-79.
Lowder. J. N., Meeker, T. C., Campbell, M., Garcia, C. F., Gralow, J., Miller,
R. A.. Warnke, R. and Levy. R. (1987) Studies on B lymphoid tumours
treated with monoclonal anti-idlotype antibodies: Correlation with
clinical responses. Blood 69,199-213.
Luan, Y. T. Mutharasan, R. and Magee, W. E. (1987a) Factors governing
lactic acid formation in long term cultivation of hybridoma cells.
Biotech. Lett. 9,761-756.
237

Luan, Y. T. Mutharasan, R. and Magee, W. E. (1987b) Strategies to extend
longevity of hybridomas In culture and promote yield of monoclonal
antibodies.
Biotech. Lett. 9.691-696.
Lubiniecki. A. S. (1987) Pharmaceutical applications of recombinant
DNA-modified mammalian cells.
Developments in Industrial Microbiology 28,
133-138.
Lydersen, B. K., Putnam, J., Bognar, E., Patterson, M., Pugh, G. G. and
Noll, L. A. (1985) The use of a ceramic matrix in a large scale cell culture
system in: Large-Scale Mammalian Cell Culture (Feder, J. and Tolbert,
W. R., eds) Academic Press, New York pp. 39-58.
Macmillan, J. C. and Wheatley, D. N. (1981) Amino acid pools In mitotic
and Interphase HeLa cells.
Exp. Cell Res. 133,470-475.
Marsani, B. D., Slate, D. L. and Schimke, R. T. (1981) S phase-specific
synthesis of dihydrofolate reductase in Chinese hamster ovary cells.
Proc. Natl. Acad. Sc!. USA 78,4985-4989.
Marston, F. A. O. (1986) The purification of eukaryotic polypeptides
synthesized in Escherichla cola.
Biochem. J. 240,1-12.
Mather, J. P. and Sato, G. H. (1979) The growth of mouse melanoma cells
in hormone supplemented, serum-free medium.
Exp. Cell Res. 120,191-200.
Mather, J., Wu, R. and Sato, G. (1981) The role of hormones In the growth
and regulation of cells in a serum-free medium in: The Growth
Requirements of Vertebrate Cells in vitro (Waymouth, C., Ham, R. G. and
Chapple, P. J., eds. ) Cambridge University Press, London pp. 244-251.
Mattlasson, B., Borrebaeck, C., Sanfridson, B. and Mosbach, K. (1977)
Thermometric enzyme-linked immunosorbent assay: TELISA.
Biochim. Biophys. Acta 483.221-227.
Maurer, H. R. (1986) Towards chemically-defined, serum-free media for
mammalian cell culture In: Animal Cell Culture, a Practical Approach
(Freshney, R. I., ed. ) IRL Press, Oxford pp. 13-32.
McCarty, K. (1962) Selective utilization of amino acids by mammalian
cell cultures.
Exp. Cell Res. 27,230-240.
McGee. R. and Spector. A. A. (1974) Short term effects of free fatty acids
on the regulation of fatty acid biosynthesis in Ehrlich ascites tumour
cells.
Cancer Res. 34,3355-3362.
McKeehan, W. L., Hamilton, W. G. and Ham, R. G. (1976) Selenium is an
essential trace nutrient for growth of W138 diploid human fibroblasts.
Proc. Natl. Acad. Sci. USA 73,2023-2027.
238

McKeehan, W. L. (1986) Glutaminolysis in animal cells in: Carbohydrate
Metabolism in Cultured Cells (Morgan. M. J., ed. ) Plenum Press pp.
111-150.
McLimans, W. F. (1972) The gaseous environment of the mammalian cell
in culture in: Growth, Nutrition and Metabolism of Cells in Culture Vol.
2 (Rothblat, G. H. and Cristofalo, V. J.. eds. ) Academic Press, London.
pp. 137-170.
Meister, A. (1962) Amide nitrogen transfer (survey).
The Enzymes 6,247-266.
Melvin, W. T., Burke, J. F. Slater. A. A. and Keir, H. M. (1979) Effect of
amino acid deprivation on DNA synthesis in BHK-21/C13 cells.
J. Cell. Physiol. 98,73-80.
Merten. O. -W.. Reiter. S.. Himmler, G., Scheirer, W. and Katinger, H.
(1985) Production kinetics of monoclonal antibodies.
Develop. biol. Standard. 60.219-227.
Merten, O. -W., Palfi, G. E. and Steiner, J. (1986) ??? In: Advances in
Biotechnological Processes, vol 6 (Mizrahi, A., ed. ) Alan R Liss, New York
pp. 111-178.
Merten, O. -W. (1987) Concentrating mammalian cells I. Large-scale
animal cell culture.
Trends in Biotechnology 5.230-237.
Milcarek, C. and Zahn, K. (1978) The synthesis of ninety proteins
including actin throughout the cell cycle.
J. Cell Biol. 79,833-838.
Miller, R. A., Maloney, D. G., Warnke, R. and Levy, R. (1982) Treatment of
B-cell lymphoma with monoclonal anti-idlotype antibody.
New Engl. J. Med. 306 , 517-522.
Miller, W. M., Wilke, C. R., and Blanch, H. W. (1987) Effects of dissolved
oxygen concentration on hybridoma growth and metabolism In
continuous culture. J. Cell. Physiol. 132,524-530.
Miller. W. M., Blanch, H. W. and Wilke. C. R. (1988a) A kinetic analysis of
hybridoma growth and metabolism in batch and continuous suspension
culture: Effect of nutrient concentration, dilution rate and pH.
Biotechnol. Bioeng. 32,947-965.
Miller, W. M., Wilke, C. R. and Blanch, H. W. (1988b) Transient responses
of hybridoma cells to lactate and ammonia pulse and step changes in
continuous culture. Bloprocess Engineering 3,113-122.
Miller, W. M., Wilke, C. R. and Blanch, H. W. (1989) Transient responses
of hybridoma cells to nutrient additions in continuous culture: 1.
Glucose pulse and step additions.
Biotechnol. Bioeng. 33,477-486.
239

Milstein, C. (1980) Monoclonal antibodies.
Sci. Am. 243 (Oct), 56-64.
Miltenburger, H. G. and Krieg, A. (1984) Bloinsecticides II. Baculovirldae
in: Advances in Biotechnological Processes, vol 3 (Mlzrahi, A. and van
Wezel. A. L., eds. ) Alan R Liss, New York pp. 291-313.
Mitchell, B. F. and Tupper. J. T. (1977) Synchronisation of mouse 3T3 and
SV40 3T3 cells by way of centrifugal elutriation.
Exp. Cell Res. 106,351-355.
Mitchison, J. M. (1971) The Biology of the Cell Cycle.
" Cambridge University Press, Cambridge.
Mlzrahl, A., Vosseller, G. V., Yagi, Y. and Moore, G. E. (1972) The effect
of dissolved oxygen partial pressure on growth, metabolism and
immunoglobulin production in a permanent human lymphocyte cell line
culture.
Proc. Soc. Exp. Biol. Med. 139,118-122.
Mizrahi, A. (1988) Biologicals produced from animal cells in culture- an
overview.
Biotech. Adv. 6.207-220.
Monod, J. (1942) Recherches sur la croissance des cultures bacteriennes.
Hermann, Paris.
Monod. J. (1950) La technique de culture continue; theorie et
applications.
Ann. Inst. Pasteur Paris 79,390-410.
Moreadith, R. W. and Lehninger, A. L. (1984) The pathway of glutamate
and glutamine oxidation by tumour cell mitochondria: Role of
mitochondrial NAD(P)+ dependent malic enzyme.
J. Biol. Chem. 259.6215-6221.
Morgan, J. F.. Morton, H: J. and Parker, R. C. (1950) Nutrition of animal
cells In tissue culture. 1. Initial studies on a synthetic medium.
Proc. Soc. Exp. Biol. Med. 73,1-8.
Morgan, M. J. and Falk, P. (1986) The utilization of carbohydrates by
animal cells in: Carbohydrate Metabolism in Cultured Cells (Morgan, M.
J.. ed. ) Plenum Press pp. 29-75.
Morris, D. L. Campbell, J. and Hornby, W. E. (1975) A chemistry for the
immoblisation of enzymes on nylon: The preparation of nylon
tube-supported hexokinase and glucose 6-phosphate dehydrogenase
and the use of the co-immobilised enzymes in the automated
determination of glucose.
Blochem. J. 147,593-603.
Morrison, S. L.. Johnson, M. J., Herzenberg, L. A. and 01. V. T. (1984)
Chimeric human antibody molecules: Mouse antigen-binding domains
with human constant region domains.
Proc. Natl. Acad. Sci. USA 81,6851-6855.
240

Morrison, S. L. (1985) Transfectomas provide novel chimeric antibodies.
Science 229,1202-1207.
Moser, H. and Vecchio, G. (1967) The production of stable steady states
in mouse ascites mast cell cultures maintained in the chemostat.
Experlentia 23,1-10.
Mueller, G. C. (1969) Biochemical events in the animal cell cycle.
Fed. Proc. 28,1780-1789.
Mueller. G. C. and Kajiwara, K. (1966) Early- and late-replicating
deoxyribonucleic acid complexes in HeLa cell nuclei.
Biochim. Biophys. Acta 114,108-115.
Murakami, H., Masui, H., Sato, G. H., Sueoka, N., Chow, T. P. and
Kano-Sueoka. T. (1982) Growth of hybridoma cells in serum-free
medium: Ethanolamine is an essential component.
Proc. Natl. Acad. Sci. USA 79,1158-1162.
Murdin, A. D., Thorpe. J. S.. Kirkby, N., Groves, D. J. and Spier, R. E.
(1987) Immobilisation and growth of hybridomas in packed beds in:
Bloreactors and Blotransformations (Moody, G. W. and Baker, P. B., eds)
Elsevier, London. pp.. 99-110.
Nakamura. H., Morita, T.. Masaki, S. and Yoshida, S. (1984) Intracellular
localization and metabolism of DNA polymerase alpha In human cells
visualized with monoclonal antibody.
Exp. Cell Res. 151,123-133.
Neurath, A. R. and Strick, N. (1981) Enzyme-linked fluorescence
immunoassays using B-galactosidase and antibodies covalently bound
to polystyrene plates.
J. Virol. Meth. 3,155-165.
Ngo. T. T. and Lenhoff, H. M. (1981) New approach to heterogeneous
enzyme immunoassays using tagged enzyme-ligand conjugates.
Biochem. Blophys. Res. Commun. 99,496-503.
Nilsson, H.. Akerlund. A. -C. and Mosbach. K. (1973) Determination of
glucose, urea and penicillin using enzyme-pH-electrodes.
Blochim. Biophys. Acta 320,529-534.
Nilsson, K. and Mosbach. K. (1980) Preparation of immobilized animal
cells.
FEBS Lett. 118,145-150.
Nilsson, K., Scheirer, W., Merten, O. -W., Ostberg, L., Liehl, E.. Katinger,
H. W. D. and Mosbach, K. (1983) Entrapment of animal cells for the
production of monoclonal antibodies and other biomolecules.
Nature 302,629-630.
Novick. A. and Szilard. L. (1950) Description of the chemostat.
Science 112,715-716.
241

Pardee, A. B. (1974) A restriction point for the control of normal animal
cell proliferation.
Proc. Natl. Acad. Sci. USA 71,1286-1290.
Pardee, A. B., Dubrow, R., Hamlin, J. L. and Kletzeln, R. F. (1978) Animal
cell cycle.
Ann. Rev. Biochem. 47,715-750.
Pardee, A. B., Coppock. D. L. and Yang, H. C. (1986) Regulation of cell
proliferation at the onset of DNA synthesis.
J. Cell Scl. Suppl. 4,171-180.
Pasieka, A. E., Morton, H. J. and Morgan, J. F. (1960) The metabolism of
animal tissues cultivated in vitro IV. Comparative studies on human
malignant cells.
Cancer Res. 20,362-367.
Petricciani, J. C. (1987) Should continuous cell lines be used as
substrates for biological products.
Develop. biol. Standard. 66,3-12. -
Phillips, A. W., Ball, G. D., Fantes, K. H., Finter, N. B. and Johnston, M. D.
(1985) Experience in the cultivation of mammalian cells on the 80001
scale in: Large-Scale Mammalian Cell Culture (Feder, J. and Tolbert, W. R.
eds. ) Academic Press, New York pp. 87-96.
Phillips, H. A., Scharer, J. M., Bols, N. C. and Moo-Young, M. (1987) Effect
of oxygen on antibody productivity in hybridoma culture.
Biotech. Lett. 9,745-750.
Prescott. D. M. and Bender, M. A. (1962) Synthesis of RNA and protein
during mitosis in mammalian tissue culture cells.
Exp. Cell Res. 26,260-268.
Puck. T. T. (1964) Phasing, mitotic delay and chromosomal aberrations
in mammalian cells. Science 144,565-566.
Radlett, P. J.. Telling. R. C., Whiteside, J. P. and Maskell, M. A. (1972)
The supply of oxygen to submerged cultures of BHK 21 cells.
Blotechnol. Bioeng. 14,437-445.
Ramabhadran. T. V. (1987) Products from genetically engineered
mammalian cells : benefits and risk factors.
Trends in Biotechnology 5,175-179.
Ratafia, M. (1987) Mammalian cell culture: Worldwide activities and
markets.
Blo/Technology 5,692-694.
Reagan, K. J., Wunner, W. H., Wiktor, T. J. and Koprowski, H. (1983)
Anti-idiotype antibodies induce neutralizing antibodies to rabies virus
glycoprotein.
J. Virol. 48,660-666.
242

Regan, D. H., Lavietes, B. B., Regan, M. G., Demopoulos, H. B. and Morris,
H. P. (1973) Glutamate-mediated respiration In tumors.
J. Natl. Cancer Institute 51,1013-1017.
Reitzer, L. J., Wice, B. M. and Kennell, D. (1979) Evidence that glutamine,
not sugar, is the major energy source for cultured HeLa cells.
J. Biol. Chem. 254,2669-2676.
Renard, J. M., Spagnoli, R., Mazier, C., Salles, M. F. and Mandine, E.
(1988) Evidence that monoclonal antibody production kinetics is related
to the Integral of the viable cells curve in batch systems.
Biotech. Lett. 10,91-96.
Renner, E. D., Plagemann, P. G. W. and Bernlohr, R. W. (1972) Permeation
of glucose by simple and facilitated diffusion by Novikoff rat hepatoma
cells in suspension culture and its relation to glucose metabolism.
J. Biol. Chem. 247,5765-6776.
Reuveny, S., Velez, D., Riske, F., Macmillan, J. D., and Miller, L. (1985)
Production of monoclonal antibodies In culture.
Develop. biol. Standard. 60,185-197.
Reuveny, S., Velez, D., Macmillan, J. D. and Miller, L. (1986a) Factors
affecting cell growth and monoclonal antibody production in stirred
reactors.
J. Immunol. Methods 86.53-59.
Reuveny, S., Velez, D., Miller. L. and Macmillan, J. D. (1986b) Comparison
of cell propagation methods for their effect on monoclonal antibody yield
In fermenters.
J. Immunol. Methods 86,61-69.
Reuveny, S., Lazar, A., Mlzrahl, A., Marcus, D., Shahar, A. and Epstein,
N. (1987a) Carcinoembryonic antigen (CEA) production in low protein
growth medium In: Modern Approaches to Animal Cell Technology (Spier,
R. E. and Griffiths, J. B., eds. ) Butterworths, London pp. 724-737.
Reuveny, S., Velez, D., Macmillan, J. D. and Miller, L. (1987b) Factors
affecting monoclonal antibody production in culture.
Develop. biol. Standard. 66,169-175.
Rhodes, M. and Birch, J. R. (1988) Large-scale production of proteins
from mammalian cells.
Blo/Technology 6,518-522.
Riechuran, L., Clark, M., Waldmann, H. and Winter, G. (1988) Reshaping
human antibodies for therapy.
Nature 332,323-327.
Ringold, G., Dieckmann, B. and Lee, F. J. (1981) Co-expression and
amplification of dihydrofolate reductase cDNA and Escherichla coil
xGPRT gene in Chinese hamster ovary cells.
Mol. Appl. Genet. 1,165-175.
243

Roberts, R. S., Hsu, H. W., Lin, K. D. and Yang, T. J. (1976) Amino acid
metabolism of myeloma cells in culture.
J. Cell Sc!. 21,609-615.
Roedere, J. E., and Bastlaans, G. J. (1983) Microgravimetric immunoassay
with piezoelectric crystals.
Anal. Chem. 55,2333-2336.
Roos, D. and Loos, J. A. (1973) Changes in the carbohydrate metabolism
of mitogenically stimulated human peripheral lymphocytes II. Relative
importance of glycolysis and oxidative phosphorylation on
phytohaemagglutinin stimulation.
Exp. Cell Res. 77,127-135.
Rous, P. and Jones, F. S. (1916) A method for obtaining suspensions of
living cells from the fixed tissues, and for the plating out of individual
cells.
J. Exp. Med. 23,549-555.
Roux, W. (1885) Beitrage zur entwicklungsmechanik
Z. Biol (Munich) 21,411-524.
des embryo.
Rubenstein, K. E., Schneider, R. S. and Ullman, E. F. (1972)
"Homogeneous" enzyme immunoassay: A new immunochemical technique.
Blochem. Blophys. Res. Commun. 47,846-851.
Rupp, R. G. (1985) Use of cellular microencapsulation in large-scale
production of monoclonal antibodies in: Large-Scale Mammalian Cell
Culture (Feder, J. and Tolbert, W. R.. eds. ) Academic Press, New York pp.
19-38.
Sadowski, P. L. (1982) European Patent Application EPO 077734A2.
Salzman, N. P., Eagle, H. and Sebring, E. D. (1958) The utilization of
Blutamine, glutamic acid and ammonia for the biosynthesis of nucleic
acid bases in mammalian cell cultures.
J. Biol. Chem. 230,1001-1012.
Schaap. G. H., van der Kamp, A. W., Ory, F. G. and Jongkind, J. F. (1979)
Isolation of anucleate cells using a fluorescence activated cell sorter
(FACS II).
Exp. Cell Res. 122,422-426.
Scharff, M. D. and Robbins, E. (1965) Synthesis of ribosomal RNA in
synchronised HeLa cells.
Nature 208,464-466.
Scharff. M. D. and Robbins, E. (1966) Polyribosome disaggregatlon during
metaphase.
Science 151,992-995.
Scheper, Th., Hoffmann, H. and Schugerl, K. (1987) Flow cytometric
studies during culture of Saccharomyces cerevislae.
Enzyme Microb. Technol. 9,399-405.
244

Schiff, S. 0. (1965) Ribonucleic acid synthesis in neuroblasts of
Chortophaga viridifasciata (de Geer) as determined by observations of
individual cells in the mitotic cycle.
Exp. Cell Res. 40,264-276.
Schliermann, M., Beckers, C. and Miltenburger, H. G. (1987) Cell cycle
analysis by flow cytometry in an antibody secreting cell line.
Develop. biol. Standard. 60,101-109.
Schlom, J. and Weeks, M. O. (1985) Important Advances in Oncology
(DeVita, V. T., Hellman, S. and Rosenberg, S. A., eds. ) J. B. Lippincott Co.
pp. 170-189.
Schonherr. 0. T., van Gelder, P. T. J. A., van Hees, P. J., van Os, A. M. M.
and Roelofs. H. W. M. (1987) A hollow fibre dialysis system for the In
vitro production of monoclonal antibodies replacing in vivo production
in mice.
Develop. biol. Standard. 66,211-220.
Schoolwerth, A. C. and Lanoue, K. F. (1980) The role of
microcompartmentation in the regulation of glutamate metabolism by rat
kidney mitochondria. J. Biol. Chem. 255,3403-3411.
Schumperli, D. (1986) Cell-cycle regulation of histone gene expression.
Cell 45,471-472.
Scott, F. M., Sator de Serrano, V. and Castellino, F. J. (1987) Appearance
of plasminogen activator activity during a synchronous cycle of a rat
adenocarcinoma cell line, PA-Ill.
Exp. Cell Res. 169,39-46.
Seaver, S. S., Rudolph. J. L. and Gabriels, J. E. (1984) A rapid HPLC
technique for monitoring amino acid utilization in cell culture.
BioTechniques 2.254-260.
Seaver, S. S. and Gabriels, J. E. (1985) A membrane-based system for
the continuous production of monoclonal antibody in culture.
Hybridoma 4,64.
Secher, D. S. and Burke, D. C. (1980) A monoclonal antibody for
large-scale purification of human leukocyte interferon.
Nature 285,446-450.
Sharath. M. D.. Rindernecht, S. B. and Weller, J. M. (1984) Human
immunoglobulin synthesis in serum-free medium.
J. Lab. Clin. Med. 103,739-748.
Shields, R., Brooks, R. F., Riddle, P. N., Capellaro, D. F. and Delia, D.
(1978) Cell size, cell cycle and transition probability in
mouse -fibroblasts.
Cell 15,469-474.
Shoham, J. (1983) Production of human immune interferon in: Advances
in Biotechnological Processes, vol 2 (Mizrahl, A. and van Wezel, A. L.,
eds. ) Alan R Liss, New York pp. 209-269.
245

Sinclair, R. (1980) Glucose metabolism and dehydrogenase activities in
the cytosol and mitochondrion of mouse LS cells in chemostat culture.
In Vitro 16,1076-1084.
Smith, J. A.. Hurrell, J. G. R. and Leach, S. J. (1978) Elimination of
nonspecific adsorption of serum proteins by Sepharose-bound antigens.
Anal. Biochem. 87.299-305.
Sonenshein, G. E. and Brawerman, G. (1976) Differential translation of
mouse myeloma messenger RNAs in a wheat germ cell-free system.
Biochemistry 15,5501-5506.
Spier. R. E. (1983) Production of veterinary vaccines in: Advances in
Biotechnological Processes, vol 2 (Mizrahi, A. and van Wezel, A. L., eds)
Alan R Liss, New York pp. 33-59.
Spier, R. E. (1983) Modern approaches to vaccines- a report of a Cold
Spring Harbor conference in modern approaches to vaccines Aug 1983.
Vaccine 1,53-54.
Spier, R. E. (1988) Animal cells in culture : moving into the exponential
phase.
Trends in Biotechnology 6,2-6.
Spier, R. E. and Horaud, F. (1985) The biotechnological future for animal
cells in culture in: Animal Cell Biotechnology, Vol 2. (Spier, R.
E., and Griffiths, J. B., eds) Academic Press, London, pp. 431-458.
Stambrook, P. J. and Sisken, J. E. (1972) Induced changes in the rates of
urldine-3H uptake and incorporation during the G1 and S periods of
synchronized Chinese hamster cells.
J. Cell Biol. 52,514-525.
Stoner, G. D. and Merchant, D. J. (1972) Amino acid utilisation by LM
mouse cells in a chemically defined medium.
In Vitro 7,330-343.
Stubblefield, E. (1968) Synchronisation methods for mammalian cell
culture In: Methods in Cell Physiology (Prescott, D. M., ed. ) Academic
Press. New York pp. 25-43.
Sutherland, R. M., Dahne, C. and Place, J. (1984) Preliminary results
obtained with a no-label, homogeneous, optical Immunoassay for human
immunoglobulin G.
Anal. Lett. 17,43-53.
Takahashi, M., Yagi, Y., Moore, G. E. and Pressman, D. (1969)
Immunoglobulin production in synchronised cultures of human
haematopoletic cell lines 1. Variation of cellular immunoglobulin level
with the generation cycle.
J. Immunology 103,834-843.
Talle, M. A., Allecar, N.. Makowski, M., Rae, P. E., Mittler, R. S. and
Goldstein, G. (1983)
Diagn. Immunol. 1,129-135.
246

Tanno, Y., Aral, S. and Takishima, T. (1982)
J. Immunol. Methods 52,255-261.
Telling, R. C. and Elsworth, R. (1965) Submerged culture of hamster
kidney cells in a stainless steel vessel.
Biotechnol. Bioeng. 7,417-434.
Terasima, T. and Tolmach, L. J. (1963) Growth and nucleic acid synthesis
in synchronously dividing populations of HeLa cells.
Exp. Cell Res. 30,344-362.
Tharakan, J. P. and Chau, P. C. (1986a) IgG production kinetics in serum
free media.
Biotech. Lett. 8.529-534.
Tharakan, J. P. and Chau, P. C. (1986b) Operation and pressure
distribution of immobilized cell hollow fiber bloreactors.
Biotechnol. Bioeng. 28,1064-1071.
Tharakan, J. P. and Chau, P. C. (1986c) A radial flow hollow fiber
bloreactor for the large-scale culture of mammalian cells.
Blotechnol. Bioeng. 28,329-342.
Tharakan, J. P., Lucas, A. and Chau, P. C. (1986) Hybrldoma growth and
antibody secretion in serum-supplemented and low protein serum-free
media.
J. Immunol. Methods 94,226-235.
Tolbert, W. R., Lewis. C., White, P. J. and Feder, J. (1985) Perfusion
culture systems for production of mammalian cell biomolecules in:
Large-Scale Mammalian Cell Culture (Feder, J. and Tolbert, W. R., eds. )
Academic Press, New York pp. 97-124.
Toole, J. J., Knoopf, J. L., Wozney, J. M., Sultzman, L. A., Buecker, J. L.,
Pittman, D. D., Kaufman, R. D., Brown, E., Shoemaker, C. Orr, E. C.,
Amphlett, G. W., Foster, W. B., Coe, M. L., Knutson, G. J., Foss, D. N. and
Hewick, R. M. (1984) Molecular cloning of a cDNA encoding human
antihemophilic
factor.
Nature 312,342-347.
Tovey, M. G., Mathison, G. E. and Pirt, S. J. (1973) The production of
interferon by chemostat cultures of mouse LS cells grown in
chemically-defined, protein-free medium.
J. Gen. Virol. 20,29-35.
Tovey. M. and Brouty-Boye, D. (1976) Characteristics of the chemostat
culture of murine leukemia L1210 cells.
Exp. Cell Res. 101,346-354.
Tovey, M. G. (1980) The cultivation of animal cells in the chemostat:
Application to the study of tumor cell multiplication.
Adv. Cancer Res. 33,1-37.
247

Trowbridge, I. S. and Omary, M. B. (1981) Human surface glycoprotein
related to cell proliferation is the receptor for transferrin.
Proc. Natl. Acad. Sci. USA 78,3039-3043.
Turner. K. J., Plozza, T. M. and Holt, P. G. (1985) Cell cycle-dependent
fluctuations in IgE secretion in a human myeloma line.
Clin. Immunol. Immunopath. 36,212-216.
Vallera, D. A., Youle, R. J., Neville, D. M. and Kersey, J. H. (1982) Bone
marrow translplantation across major histocompatibill ty barriers V.
Protection of mice from lethal graft-vs-host disease by pretreatment of
donor cells with monoclonal anti-Thy-1.2 coupled to the toxin ricin.
J. Exp. Med. 155,949-954.
Van Hemert. P., Kilburn, D. G. and van Wezel, A. L. (1970) Homogeneous
cultivation of animal cells for the production of virus and virus
products.
Biotechnol. Bioeng. 11,875-895.
Van Wauwe, J. and Goosens, J. (1981) Monoclonal anti-human
T-lymphocyte antibodies: Enumeration and characterisation of T-cell
subsets.
Immunology 42,157-164.
Van Wezel, A. L., van der Marel, P., van Beversen, C. P., Verma, I., Salk,
P. L. and Salk, J. (1982) Detection and elimination of cellular nucleic
acids in biologicals produced on continuous cell lines.
Develop. biol. Standard. 50,59-69.
Van Wezel, A. L. van der Velden de Groot, C. A. M., de Haan, H. H., van
den Heuvel, N. and Schasfoort, R. (1985) Large scale animal cell
cultivation for production of cellular biologicals.
Develop. biol. Standard. 60,229-236.
Velez, D.. Reuveny, S., Miller, L. and Macmillan, J. D. (1986) Kinetics of
monoclonal antibody production in low serum growth medium.
J. Immunol. Methods 86,45-52.
Visek, W. J., Kolodny, G. M. and Gross, P. R. (1972) Ammonia effects in
cultures of normal and transformed 3T3 cells.
J. Cell. Physiol. 80,373-382.
Vitetta, E. S., Fulton, R. J.. May, R. D., Till, M. and Uhr, J. W. (1987)
Redesigning natures poisons to create anti-tumor reagents.
Science 238,1098-1104.
Voak, D. and Lennox, E. S. (1983)
Biotest Bulletin 4,281-290.
Warburg, 0. and Kubowitz, F. (1927) Stoffwechsel wachsender zellen.
Biochem. Z. 189,242-248.
Warburg, 0. (1930) The metabolism of tumours. Constable, London.
248

Warburg, 0. (1956) On the origin of cancer cells.
Science 123,309-314.
Watanabe, S., Yagi, Y. , and Pressman, D. (1973) Immunoglobulin
production in synchronised cultures of human hematopoletic cell lines
II. Variation of synthetic and secretion activities during the cell cycle.
J. Immunol. 111,797-804.
Waymouth, C. (1981) Requirements for serum-free growth of cells:
Comparison of currently available defined media in: The Growth
Requirements of Vertebrate Cells in vitro (Waymouth, C., Ham, R. G. and
Chapple, P. J.. eds. ) Cambridge University Press, London pp. 33-47.
Weetall, H. H. (1976) Covalent coupling methods for Inorganic support
materials in: Methods in Enzymology, vol 44. (Jakoby, W. B. and Pastan,
I. H., eds. ) Academic Press, New York pp.
Westerwoudt. R. J. (1986) Factors affecting production of monoclonal
antibodies in Methods in Enzymology. vol 121. (Langone. J. and van
Vunakis, H. eds. ) Academic Press, London pp. 3-18.
Westwood, J. T., Wagenaar, E. B. and Church, R. B. (1986) Synthesis of
unique low molecular weight proteins during late G2 and mitosis.
J. Biol. Chem. 260,695-698.
White, P. R. (1949) Prolonged survival of excised animal tissues in
nutrients of precisely known constitution.
J. Cell Comp. Physiol. 34,221-241.
Wice, B. M., Reitzer, L. J. and Kennett, D. (1981) The continuous growth
of vertebrate cells in the absence of sugar.
J. 8101. Chem. 256,7812-7819.
Wiemann, M. C., Ball, E. D., Fanger. M. W., Dexter, D. L., McIntyre, 0. R.,
Brnier, G. and Calabresi, P. (1983) Human and murlne hybridomas: Growth
and monoclonal antibody production in the artificial capillary culture
system.
Clin. Res. 31,611A.
Wiktor. T. J. and Koprowski, H. (1978) Monoclonal antibodies against
rabies virus produced by somatic cell hybridization: detection of
antigenic variants. Proc. Natl. Acad. Scl. USA 75,3938-3942.
Williams, D. C., Van Frank, R. M.. Muth, W. L. and Burnett, J. P. (1982)
Cytoplasmic inclusion bodies In Escherichla cola producing biosynthetic
human insulin proteins.
Science 216,687-688.
Wilson, M. B. and Nakane. P. K. (1976) The covalent coupling of proteins
to periodate-oxidised Sephadex: A new approach to immunoadsorbent
preparation.
J. Immunol. Methods 12,171-181.
249

Wood, C. R., Boss, M. A., Kenten, J. H., Calvert, J. E., Roberts, N. A. and
Emtage, J. S. (1985) The synthesis and In vivo assembly of functional
antibodies in yeast. Nature 314,446-449.
Wood. W. I., Capon, D. J., Simonsen, C. C., Eaton, D. L., Glteschier
, If.,
Keyt, B., Seeburg, P. H., Smith, D. If., Hollingshead, P.,, Wion, K. L.,
Delaware, K., Tuddenham, G. D., Vehar, G. A. and Lawn, R. M. (1984)
Expression of active human factor VIII from recombinant DNA clones.
Nature 312,330-337.
Wright, A. F., Green, T. P. and Smith, L. L. (1987) The development of
mouse monoclonal antibodies that bind and neutralize paraquat.
Develop. blol, Standard. 66,496-601.
Wynford-Thomas, D., Lamontagne, A., Marin, G. and Prescott, D. M.
(1985) Location of the isoleucine arrest point In CHO and 3T3 cells.
Exp. Cell Res. 168,625-632.
Xeros, N. (1962) Deoxyriboside control and synchronisation of mitosis.
Nature 194,682-683.
Yabe, N., Kato, M., Matsuya, Y., Yamane, I., Izuka, M., Takayoshl, H. and
Suzuki, K. (1987) Role of iron chelators in growth-promoting effect on
mouse hybridoma cells in a chemically defined medium.
In Vitro 23,816-820.
Youle, R. J. and Neville, D. M. (1980) Anti-Thy-1.2 monoclonal antibody
linked to ricin is a potent cell-type-specific toxin.
Proc. Natl. Acad. Scl. USA 77,6483-6486.
Zetterberg, A. and Engstrom, W. (1981) Glutamine and the regulation of
DNA replication and cell multiplication In fibroblasts.
J. Cell. Physiol. 108,365-373.
Zetterberg, A. and Killander, D. (1965) Quantitative cytophotometric and
autoradlographic studies on the rate of protein synthesis during
interphase in mouse fibroblasts In vitro.
Exp. Cell Res. 40,1-11.
Zielke, H. R., Ozand, P. T., Tildon, J. T., Sevdallan, D. A. and Cornblath,
M. (1978) Reciprocal regulation of glucose and Blutamine utilization by
cultured human diploid fibroblasts.
J. Cell. Physiol. 95,41-48.
Zlelke, H. R., Sumbilla, C. M., Sevdallan, D. A., Hawkins, R. L. and Ozand,
P. T. (1980) Lactate: a major product of glutamine metabolism by human
diploid fibroblasts.
J. Cell. Physiol. 104,433-441.
Zimmerman, U., Vienken, J., Halfmann, J. and Emels, C. C. (1986)
Electrofuslon, a novel hybridization technique In: Advances In
Biotechnological Processes, vol 4 (Mizrahl, A. and van Wezel, A. L., eds. )
Alan R Liss, New York pp. 79-94.
260

NTT
t, 2 V;;
endie Sole od 113terface.
Solenoid Interface
12V Power Supply
Solenoid Switching Unit
251
I
regulator 7812
4700uF 0.22 0.47
solenoids
(00006
1ý
+12V
ov
a. R kn
v led'"

Appendix II. Automated ELISA control program.
10 REM P. M. Hayter 1/11/85
20 REM Online assay control program
30
40
50 REM Set variables. Dimension arrays-Set up screen
60 no% i
70 DIM data(200)tDIM number%(100)tDIM time%(100)sDIM reading(100)
80 MODE4, COLOUR 131tCOLOUR 0
90
100 REM Load graphics dump routine
110 *LOAD"DUMP"
120
1ti0 REM Set all lines on user port for output
140 PROCuser
150
160 REM Set up spectrophotometer
170 PROCspec
180
190 REM Start routine
200 PROCdetails
210 PROCfi1e
220 PROCgo
230
240 REM Control routine to switch solenoid valves and assay sample
250 ON ERROR GOTO 326J
260 PROCsampletF'ROCabsorb
270
280 REM Calculate peak area and store data
290 PROCvalue
300 PROCsavetPROCdelay(2)
310
320 REM Update graph showing progress of the culture
330 FROCplot: PROCelute
340
350 REM Return to start of sequence
360 no%-no%+i: GOT0260
370
380
390 DEFFROCspec
400 PROCcalibrate
410 PROCsetup
420 ENDFROC
4Z0
440
450 DEFPROCsutup
460 VDU121PRINT 7Aß(6.10)"Setting up spectrophotomettr"IFRINT TAE'1ýý12)"Flýeat
e wait"
470 PROCinitialisesFRCCdelav(2)
480 commandt="W. 405"
490 FROCsend(commandt)sF'RDCdelay(20)
500 command$-"REF. "tPRCCsend(commandt)%PROCdolay(1)
510 ENDFROC
520
530
540 DEFPROCinitialise
550 "FX7,2
560 *FX8,2
570 *FX2,0
S8') *FX166,1,252
590 ENDF*RCC
252

-60o
610
620 DEFF'ROCsend(command. t)
670 "FX:. 7
640 PRINT command$; CHR$(10)i
66O "FX3, O
660 ENDPROC
670
680
690 DEFPROCdelay(interva1%)
700 time7. TIME+(100"intervwl%)
710 REPEAT
720 UNTIL TIME>time%
730 ENDFROC
740
750
760 DEFPROCwrvia(address%. valuv%)
770 address%-address% AND &FF
780 IF (address%&6F) THEN ENDFROC
790 A%n&973X%-address%tY%wvalue'I. tCALL &FFF4
800 ENDPROC
810
820
8Z0 DEFF'ROCuser
840 F'ROCwrvia(L
11: 0 DEFF'ROCf ile
1140 PRINT TAB(1.20)"Enter filename fo" data storage "tINPUT TAB(16,22)f1let
1160 IF LEN(filei)>6 THEN VDU7sPRINT TAB(16,22)" "tGOTO 1140
1100 IF filet-""THEN ENDPROC
1170 ON ERROR GOTO 1240
1180 channel-OPENIN filet
1190 INFUT£channel, ES, E$, ES, ES
1200 CLOSE channel
1210 PRINT TAS(1,20)"** FILE IN USE-ENTER NEW NAME "*"1PRINT TA8(16.22)"
1220 PROCdeley()
1250 GOTO 1140
1240 IF EFR222 THEN REFORTsON E'nF0 OF=iGOTO 210
263

1250 ON EKRUR OFF
1260 channel-OF"ENOUTfilet
1270 PRINTEchannel, datet. timet. meds. dil$. cell*
1280 GOTO 220
1290
1300
1310 DEFPRDCgo
1320 VDU12sINFUT TAS(5910)"Press RETURN to start sampling "AS
1330 IFA$1"I THEN VDU7sGOT0 1320
13 0 VDU12iT%=TIME
1360 ENDPROC
1370
1380
1390 DEFPROCsample s
1400
1410 REM Sample from vessel
1420 PRINTsFRINT TAB(15)"Sampling vessel"
1430 PROCwrvia(&60. &C3)
1440 PROCdelay(SO)
1450
1460 REM Sample onto column
1470 VDU12iPRINTsPRINT TA6(12)"Sample onto column"
1480 PROCwrvia(&60. &20)
1490 PROCdelay(30)
1200
1510 REM Conjugated antiserum onto column
1520 VDUI2IPRINTIPRINT TAS(10)"Antiserum onto column"
15; 0 PROCwrvia(&60. $. 27)
1540 PROCdelay(20)
1550
1560 REM Buffer wash
1570 VDU12sF'RINTtPRINT TAB(1S)"Suffer wash"
1630 PROCwrvia(&60. &20)
1690 PROCdelay(Z. 0)
1600
1610 REM Substrate onto column
1620 VDUI2zPRINTsPRINT TAE(12)"Substrate onto column"
1630 PROCwrvia(&60. &28)
1640 PROCdelay(45)
1650
1660 REM React for 10 minutes
1670 VDUI2iPRINTsPRINT TAb(16)"Substrate reaction"
1680 PROCwrvia(&60. &A0)
1690 PROCdelay(Z)
1700
1710 REM Buffer on
1720 VDUI2sPRINTsPRINT TA9(15)"Reading absorbance"
1730 PROCwrvia(&60, L

1920 8s=M1Dz'AS, 14, s)sdata(l)"VAL(B$)
19-'0 *FX3,0
1940 *FX2,0
1950 ENDPRCC
1960
1970
1950 DEFPROCvalue
1990 a-Otmax-0
2000 FOR I-1TOS0
2010 a-data(I)
2020 IF a>max THEN max-atpeak%-I
2030 NEXT
2040 FOR Impeak%-5 TO peak'!. -lsmax-max+data(I)sNEXT
2050 FOR I=peak;: +1 TO peak7.. +5smax max+dataýI)sNEXT
2060 reading(no%)-maxtr-max
2070 ENDPROC
2080
2090
2100 DEFPROCsav!
2110 PRINT"Saving data"
2120 time%(no%)-TIME-T%tt . time;: (no'l. )
211-0
2140
n%. -no':
PRINT£channel. n`/., t%, r
2150
2160
2170
ENDF"ROC
2180
2190
DEFPRCCelute
22t: 0 REM Elute with 3M NaSCN
2210 PRINTsPRINT TAB(15)"Regenerating column"
2220 PROCwrvia(&60, &. 3)
22,0
2240
PRCCdelay(180)
2250
2260
2270
REM Wash with buffer
PRINTtPRINT TAB(15)"Buffer
PROCwrvia(w60. &20)
wash"
2280 PROCdelay(600)
2290
2300
2310
ENDPROC
2320 DEFF"ROCmenu
23.30 CLOSE£0
2.3.40 VDU26.12iF"RINT TAB(2,5)"1. Ftle catalogue"
2350 PRINT TAB(2,6)"2. Recall data"
2360 PRINT TAB(2,7)"3. Plot data"
2370 PRINT TAB(2,8)"4. Print graph"
2380 PRINT TAB(2,9)"S. Restart program"
2390 PRINT TAB(2,10)"6. Exit program"
2400 PRINT TAB(2,20)"Current files"; filet
2410 INPUT TAB(5,15)"Which"IA
2420 ON A GOTO 24.30,2450,2470,2490,2600,2510
2470 VDU121*.
2440 PRINT TAB(5,20)"Press return for menu"sINFUT8IsIF D$-"" THEN 2340
2450 PROCrecall
2460 PRINT TAB(6920)"Press return for menu"sINPUTb$1IF 6i "" THEN 2340
2470 PROCplot
2480 PRINT TAB(5.1)"Press return for menu" iINPUTBtt IF @ts"" THEN 21.40
2490 PROCprinttGOTO 2340
2500 PROCrestartsENDPROC
2510 VDUI2sPRINT TAB(20,10)"** END OF PROGRAM **"LEND
2520 ENDF"ROC
2530
2540
2550 DEFF'ROCrecal1
2560 ON ERROR GOTO 3330
2570 LOCAL Ill-l
265

2580 VDUI2: PRINT TAr(14.10)"Enter filename": INPUT TAB(18.12)stcre$
2--90 IF LEN(storeS)>6 THEN VDU7: GOTO2530
2600 IF stores="" THEN ENDFGOC
2610 channel=CPENIldstoret
2620 INF, UT£channel. date$, timei, med1, dil$, cell3
2630 REPEAT: INPUT£channel, number%(I), time%(I), reading(I): I=I+1: UNTIL EOF£channe
1
2640 CLOSE£channel: no'%=number%(I)SPRINT TAE+(10,15)"Data loaded"
2650 ENDPROC
2660
2670
2680 DEFPROCplot
2690 PROCaxes
2700 PROCdraw
2710 ENDPROC
2720
2730
2740 DEFPROCaxes
2750 VDU28,0,31; 39,28: GCOL0,128: GCOL0,3: VDU16,12
2760 MCVE260,31C: DRAW! 260,313: DRAW1260,1013sDRAW260,1013: DRAW260,313
2770 FOR I=260TO1260STEP200: MOVE I, 313: DRAW I, 3Z0: MOVE I, 996; DRAW I, 1013: NEXT
2780 FOR I=313TDIOIZSTEP140: MOVE260, I: DR.; W277,1: MOVE1260, I: DRAW124'-, I: NEXT
2790 tmax%=(time'l. (no'! ))/360000: IF tmax%

3210
32-10
3240
3250
3260
3270
3280
3: QV
33100
3310
3320
=30
3340
3350
3360
3270
3x80
3390
3400
3410
DEFFR, OCrestart
no%=ä
ENDF"ROC
REM *** Escape routine ***
IF ERR17 THEN 3320
VDU26,12
PRCCmenu
PROCelute
ON ERROR OFF: GOTO 200
PRINT TAB(2.10)"Error "; ERR; " at line "; ERL: REPORTzGOTO 3280
REM *** Disc errors ***
IF ERR=222 THEN E. $="**FILE NOT ON DISC-": E1$="ENTER AGAIN**"sGOTO 3400
IF ERR=190 THEN ES="**CATALOGUE FULL-' GOTO 3390
IF
IF
ERR=191
ERR=198
THEN ES="**DISC
THEN ES '**DISC
FULL-'aGOTO
FULL-"
3390
4.
IF ERR198 THEN REPORT: PRINT ERR: PRINT ERL%GOTOZ41O
E1*="USE NEW DISC**"
VDU7: PRINT TAB(2.1O)Et;: PRINT E13
PROCdelay(5): ON ERROR OFF300TO 3290
257

Appendix III Cross-flow filter
iý
ococo
o- 4==b
oo c==p
o co 0o cý
00ov
I==
Filtrate
258
i
Silicone 0-ring
0.22 pm fitter
Fitter support
Medium
flow

Appendix IV. The determination of the volumetric mass transfer
coefficient (KL a) for oxygen in the Gallenkamp fermenter.
The volumetric mass transfer coefficient for 02 can be calculated
by depleting the fermenter of oxygen by flushing with nitrogen and
observing the rate of oxygen transfer when the oxygen supply is
restored.
The relationship between K,. a and the rate of oxygen transfer is
described by the following equation:
-KLat = ln(C, at - Ct)
where Cs at is the oxygen concentration at air saturation and Ct is the
oxygen concentration at time t. Thus Ki. a can be derived from a plot of
ln(Cs at - Ct) versus time where -KL a is equal to the slope of the
resulting graph.
The graph below shows the determination of KL a for the
Gallenkamp fermenter at stirrer speeds of 35 and 100 rpm. The fermenter
contained 500m1 RPMI 1640 and the headspace was gassed with air at
3litres/h. These determinations were carried out at 370 C using a
Gallenkamp galvanic oxygen probe.
259

u
ä
2.0
1-8
1.6
1.4
1.2
1.0
0.6
0.4
0.2
0
-0.2
-0.4
Kea determination in the Gallenkamp
fermenter
260
U IVERSf1Y
OF
SUB
BEY
Lil it
2.5