12_ANIMAL CELL TECHNOLOGY_707

USE OF A NEW MICROCARRIER WITH TWO-DIMENSIONAL
GEOMETRY FOR THE CULTURE OF ANCHORAGE-DEPENDENT CELLS

IN SONOPERFUSED, CONTINUOUSLY STIRRED TANK REACTORS

C. GATOT(1), F. TRAMPLER(2), M.-P. WANDERPEPEN(1),
J. HARFIELD(3), A. OUDSHOORN(4), A. JOHANSSON(5),
V. NIELSEN(5), A.O.A. MILLER(6)
(1) Computer Cell Culture Center s.a. (Mons, Belgium) ; (2) Sonosep
Biotech Inc (Vancouver, Canada) ; (3) Coulter Electronics Ltd (Luton,
UK) ; (4) Applikon Dependable Instruments (Schiedam, Holland) ;
(5) Nunc A/S (Roskilde, Denmark) ; (6) Biochimie/Technologie Cellulaire

- Faculté de Médecine - Université de Mons-Hainaut (Mons, Belgique)

Dextran-based microbeads characterized by a 3-dimensional geometry (3D
microsupport), on the surface of which anchorage-dependent cells (ADC's) attach
and multiply, constitute a reference for the large scale cultivation of this type of cells.
Using commercially available 3D microsupport at 5 g/L, allows one to reach cell
concentrations around cells/ml, some 7 % of the culture volume being
occupied by the microcarrier.
Growing ADCs to higher cell densities can be obtained by increasing the
concentration of the microsupport. However there rapidly conies a moment when the
volume of the swollen microbeads becomes so high as to represent an important
proportion of the culture volume, decreasing accordingly the volume of the growth
medium available to the cells.
Macroporous microbeads accomodate more cells per unit volume but their highly
convoluted inner structure makes quantitative release of cells very difficult without
extensive trypsinisation.
Recently, a new generation of microcarriers with 2D geometry has been made
available for evaluation. Photo 1 gives the rationale underlying their manufacture. It
shows that the outside surface of the microbead available for cell growth
equals the sum of the surfaces of two superimposed very thin discs situated at the
equator. However the combined volumes of two such discs which becomes smaller
the thinner the film used to manufacture them (Blodgett's monomolecular films
would be ideal in this respect), represent only a very small fraction of the volume of
the corresponding sphere (2D microhexagons cut from 25 microns thick polystyrene
films have been preferred to discs since their manufacture generates less wastes
(Photo 2).
The characteristics of the 2D microhexagons used in this study are given in
Table I.

377

O.-W. Merten et al. (eds.). New Developments and New Applications in Animal Cell Technology, 377-380.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

378
Photographs 3 and 4 give the growth kinetics of mouse L929- and Muntjac cells

respectively.

379

Additional results also obtained from batch cultures conducted this time at
at confluency) with 2D microhexagons suggest that still
higher cell concentrating could be obtained provided the surface made available to
cells is further increased and the growth medium renewed by perfusion.
Acoustic perfusion is a re cently developed technology whereby isolated cells in
suspension can be grown at concentrations of and higher. In this
system, isolated cells in suspension accumulate at the nodes of a resonating acoustic
wave. At regular intervals, in order to empty the resonating chamber, the
accumulated cells are aspirated back into the culture by means of the suspension
culture continuously recirculating at the apex of the resonating chamber at twice the
perfusion rate.
To adapt this technology to the cultivation of ADC's, recirculation is strictly
prohibited since it would provoke extensive detachment of the cells from the
microsupport.
Sonoperfusion at 4L/day (4 culture volumes/day), of a 1L culture containing
naked 2D microhexagons (8,75 g/L) shows few if any microsupport particles
reaching the acoustic chamber. This effect is interpreted as being the result of the
long (± 14 cm) large bore tube bringing the microcarrier suspension to the transducer
playing the role of a settling tube.
To compensate for the observed decreased density of cell-laden 2D
microhexagons, a culture of 750 ml (inoculated with 7,5 g of microcarrier and

380

) is perfused at l,5L/day only (2 culture volumes/day). Three
days after inoculation the cell concentration reaches allowing the
collection of close to the maximum theoretical value of
This 91,5 % recovery is obtained without recourse to trypsinisation by simply forcing
several times the cell-laden 2D microhexagons suspension in 0,2 % EDTA-PBS
through the orifice of a 25 ml capacity pipette.
Provided means are found to prevent cell-laden 2D microhexagons from escaping
through the resonating acoustic wave when very high perfusion rates are used, our
results suggest that cell concentrations of can be routinely obtained
with the microsupport occupying only some 30 % of the total culture volume, making
it envisageable to reach even higher cell concentrations in the future.

Discussion Do you see aggregation of the carriers in the acoustic field or later
Anon: in the reactor?

Miller: The microcarriers-carriers align perfectly at the nodes of the wave
and when they are returned back to the suspension they are totally
isolated.

Ozturk: What limits the cell density: just cell retention, particle number,
oxygenation, or microcarrier concentration?

Miller: We have enough microcarrier area to support 25 million cells/ml,
theoretically.

Ozturk: Do you sparge?

Miller: No, we have tried horizontal semi-permeable tubing but the micro-
carriers sick horizontally. So now we are using a Braun device
which is placed vertically.

Aunins; Comparing the spherical and flat carriers, you need twice as many
particles for the same surface area and they need inoculating on
both sides. Do the inoculum requirements practically limit the

multiplication ratios and scale-up?

Miller: Our work is so preliminary that I cannot really answer your
question. Starting in batch culture, we are reasonably confident

that we can reach the theoretical results which we anticipate.

EVALUATION OF POROUS MICROCARRIERS IN FLUIDIZED
BED REACTOR FOR PROTEIN PRODUCTION BY HEK 293
CELLS

M.A. VALLE1,2, J. KAUFMAN1, W.E. BENTLEY2, J. SHILOACH1

1. Biotechnology Unit, NIDDK, NIH
Bethesda, MD 20892 USA
2. Center for Agricultural Biotechnology, University of Maryland
Biotechnology Institute, University of Maryland
College Park, MD 20742 USA

1. Purpose of Study

In an effort to obtain large quantities of the human parathyroid

binding domain, a recombinant protein secreted extracellularly in very

low quantities from the anchorage dependent HEK 293 cells, polyethylene-

silica (Cytoline ) macroporous microcarriers in a fluidized bed reactor

(Cytopilot- ) were examined.

2. Introduction

Microcarriers allow the cultivation of anchorage-dependent cell lines with
efficient use of bioreactor volume. For sensitive cell lines, macroporous mi-
crocarriers are more advantageous. The pores protect the cells from shearing
and because of the larger available surface area, the carriers can hold high
cell densities (Nilsson et al., 1986)(Shiragami et al., 1993).

Fluidized bed reactors (FBR) are a good complement to macroporous
microcarriers (Kratje and Wagner, 1992). In addition to providing a ho-
mogeneous environment under gentle mixing, they offer good mass transfer
characteristics, are easy to monitor allow for high cell densities and are
readily scaled up. Of special interest is the modular FBR, which eliminates
the necessity of a recirculating loop with its associated handling problems
(Reiter et al., 1991).

Human embryonic kidney cell line 293 (HEK 293) is a good candidate
for culture with macroporous microcarriers in FBRs. As the name indicates,

381

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 381-384.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

382

these cells are derived from secondary embryonic kidney cells transformed
with sheared fragments of adenovirus (Graham et al., 1977). They are an-
chorage dependent and easily transformed to produce heterologous proteins
(Berg et al., 1993)(Hamilton et al., 1993). Few attempts have been made to
characterize this cell line and define the optimal conditions for its growth
and protein production. This knowledge is essential if future scale-up is to
be attempted.

3. Materials and Methods

3.1. CELL LINES AND MEDIA

Human embryonic kidney cells (HEK 293) were obtained from ATCC (CRL

1573) (USA), and were routinely passaged in tissue culture flasks (Costar

Corporation, USA) using DMEM (Biofluids Inc., USA) supplemented with

10% fetal bovine serum (Biofluids). HEK 293 transformed to produce the

extracellular domain of the human parathyroid receptor (HEK 293add)

were obtained from NPS Pharmaceuticals (USA) and passaged in T-flasks

using DMEM containing 10% FBS and 200 u/mL Hygromicin B (Cal-

biochem Corporation, USA). Production medium for HEK 293add con-

sisted of DMEM containing 293 Serum-Free Supplement (Shiloach et al.,

1996), 2 mM glutamine (Gibco, USA), 200 U/mL Hyg B and 10,000 U/mL

Penn-Strep (Gibco) .

3.2. REACTOR SYSTEM AND CULTIVATION CONDITIONS

Stationary experiments were conducted in tissue culture flasks

(Costar). Cells in mid-exponential to late exponential growth phase were

inoculated at a density of and incubated at 37°C in

5% atmosphere.

Bioreactor experiments were conducted in a 2L modular fluidized bed

reactor (Cytopilot- ) (Pharmacia, Sweden) containing 300-400 mL

(15-20% v/v) macroporous microcarriers (Cytoline , Pharmacia). pH

and dissolved oxygen concentration were kept at values of 7.00 and 30%

saturated air respectively sparging with and .

3.3. ANALYSIS OF SAMPLES

Cell density was determined using a Coulter Counter ZM (Coulter Electron-
ics, FRG).Glucose, lactate and glutamine concentrations were determined
using a YSI analyzer (Yellow Spring Instruments, USA). Ammonia con-
centration was determined using a Biolyzer (Eastman Kodak Corporation,
USA).

383

4. Results and Discussion

For evaluating the system, fermentations using HEK 293 and HEK 293add

were conducted in the Cytopilot- and T-flasks. Results from these

fermentations are shown in Table 1.

In the first 100 hours after attachment, HEK 293 cells in the Cytopilot-

showed exponentially increasing consumption of glucose and pro-

duction of lactate. After 100 hours, the rates of glucose and glutamine con-

sumption and lactate and ammonia production became constant indicating

that stationary or maintenance phase was reached.

The fermentation of HEK 293add in the Cytopilot- did not per-

form as smoothly as the HEK 293 fermentation due to interruptions caused

by temperature and dissolved oxygen control problems. Nevertheless, dur-

ing the first 70 hours after attachment, the cells exhibited exponential con-

sumption of glucose and lactate production, with faster rates than the ones

observed with the HEK 293 cells and a smaller yield of lactate on glucose.

Rates of consumption of glutamine and production of ammonia during this

phase were faster with the HEK 293add than with the HEK 293 and yield

of ammonia on glutamine was also larger. After 330 hours of operation,

the growth medium was replaced with production medium. Constant rates

384

of glucose consumption and lactate production indicated a maintenance

phase. These rates were slower than the ones observed in the HEK 293

fermentation during maintenance phase even though yield of lactate on

glucose was similar.

The yield constants in the T-flasks were similar to the yield constants

observed in the fermentor. However, the culture was not allowed to reach

stationary phase.

The fermentations in the Cytopilot- were conducted in repeated

batch mode as opposed to perfused mode as suggested by the manufacturer.

Some problems resulted from this decision. The small head space in the

reactor did not allow efficient gas exchange and the filters often became

clogged. Sampling was made difficult because the volume of medium that

can be removed without affecting recirculation is small. Also, the process

of medium replacement puts the beads under violent bubbling and mixing

which might have a detrimental effect on the cells.

References

Berg, D.T., Mcclure, D.B. and Grinnel, B.W. (1993) High-level expression of secreted
proteins from cells adapted to serum-free suspension culture, Biotechniques 14, 972-
978.
Graham, F.L., Smiley, J., Russel, W.C. and Nairn, R. (1977) Characteristics of a human
cell line transformed by DNA from human adenovirus type 5, J. gen. Virol. 36, 59-72.
Hamilton, B.J., Lennon, D.J., Im, H.K., Seeburg, P.H. and Carter, D.B. (1993) Stable
expression of cloned rat GABA-A receptor subunits in a human kidney cell line,
Neuroscience Lett. 153, 206-209.
Kratje, R.B. and Wagner, R. (1992) Evaluation of production of recombinant human
interleukin-2 in fluidized bed bioreactor, Biotechnol. Bioeng. 39, 233-242.
Nilsson, K., Buzsaky, F. and Mosbach, K. (1986) Growth of anchorage-dependent cells
on macroporous microcarriers, Bio/Technology 4, 989-990.
Reiter, M., Blml, G., Gaida, T., Zach, N., Unterluggauer, F., Doblhoff-Dier, O., Noe,
M., Plail, R., Huss, S., and Katinger, H. (1991) Modular integrated fluidized bed
bioreactor technology, Bio/Technology 9, 1100-1102.
Shiloach, J., Kaufman, J., Trinh, L. and Kemp, K. (1996) Continuous production of
the extracellular domain of recombinant receptor from HEK 293 cells using
novel serum free medium, in Carrondo, J.T., Griffiths, B. and Moreira, J.L.P. (eds.),
Animal Cell Technology, Kluwer Academic Publishers, Dordrecht, pp. 535-540.
Shiragami, N., Honda, H. and Unno, H. (1993) Anchorage-dependent animal cell culture
using a porous microcarrier, Bioproc. Engin. 8, 295-299.

FLUIDIZED BED TECHNOLOGY: INFLUENCE OF FLUIDIZATION
VELOCITY ON NUTRIENT CONSUMPTION AND PRODUCT EXPRESSION

G. BLÜML1, M. REITER2, TH.GAIDA2, N. ZACH2, A. ASSADIAN2,
C. SCHMATZ2 and HERMANN KATINGER2
1)Pharmacia Biotech Europe, Vienna, Austria
2)Institute for Applied Microbiology, Vienna, Austria

Introduction
Fluidized beds in combination with macroporous microcarriers are based on the
perfusion technology. Perfusion technology was developed as it was recognised that
cells in vivo are continuously supplied with blood, lymph, or other body fluids to keep
them in a constant physiological environment. To exploit the high productivity
potential per reactor volume of a fluidized bed system optimisation of fluidization
velocities for anchorage dependent as well suspension cells immobilized in
macroporous matrices are needed to guarantee a sufficient nutrient supply similar to in
vivo conditions.

Fluidized bed with internal circulation
Fig 1 shows the principle of the internal circulation with microsparging for
oxygenation. In the Cytopilot (Vogelbusch / Pharmacia Biotech) oxygen microbubbles
are sparged homogeneously into the downcomer flow of the draft tube and uniformly
distributed by the impeller (Fig. 1). This minimizes on one hand pO2 gradients of the
system and on the other hand increases theoretical height of the fluidized bed
enormously.

385
O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 385-387.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

386
Culture Run Comparison
During the Fermentation run with Cytoline 2 a bed
expansion of 120 % (fluidization velocity 12 cm/min)
lead to a concentration gradient of nutrients throughout
the carrier bed. The Volume stream was 2.1 L/min that
equals one medium exchange in 8.3 min, which is
insufficient for high density cultures. Fig 2a shows Mol
produced lactate per Mol consumed Glucose. The
Quotient was 50% higher than the CSTR and
FBCytolinel which were well supported with Oxygen.
After increasing the fluidization velocity to 25 cm/ min
the culture of FBCytoline2 came nearly to the same
level of lactate production per Glucose consumption.
Fig 2 b (showing the relation between secreted product
per consumed Glucose) demonstrates the same effect of
circulation rate on sufficient nutrient supply. After

increasing the circulation rate to 25 cm/min (day 57) the culture behaved similar to
CSTR and fluidized bed culture with heavier beads. 2 Amino acids (Arginin and
Serin) were analysed at different Glucose levels on all 3 culture runs. The CSTR and
FBCytolinel (45 cm/min) are much more similar in spite of the different fermentation
system than the slow circulating culture FBCytoline2 (Fig 2c,d)

Summary
These fermentation runs showed that the minimum bed expansion for light beads like
Cytoline 2 is 200% at 25cm/min to give a reasonable nutrient supply to the cells in the
pores. For Heavier beads like Cytoline 1 a bed expansion of 150% at 45 cm/min is
enough. That implements a higher carrier load in the fermenter than with lighter

387
beads. The disadvantage of the higher speed is the higher shear stress especially for
sensitive cells like hu hybridoma cells in the attachment phase of the culture.
The relation between volume specific productivity of hu-hybridoma and fluidization
velocity shows an optimum of about 25 cm/min using Cytoline2 (sufficient medium
supply and low shear stress). Using Cytolinel, CHO cells (expressing the same
monoclonal antibody) and a fluidization rate of 45 cm/min (necessary for oxygen
supply) the productivity could be increased by a factor of 3 compared to the fluidized
bed with hu-hybridoma and by a factor of 30 compared to the CSTR (Tab 2).

The required fluidization velocity in Tab 3 are 3-4 times lower than the sedimentation
velocities of Cytoline 1 or 2. There is a small risk of flushing out the carriers from the
upper chamber of the fluidized bed.

Product formation could be growth dependent for some cell lines. In that case a higher
fluidization velocity could be useful to create cell bleeding out of the pores to keep the
growth rate as high as possible. The fluidization velocity could be reduced about 15 -
20 % by using macromolecules such as Dextran, Pluronic, Xanthan etc., maintaining
the same bed expansion.

HIGH DENSITY AND SCALE-UP CULTIVATION OF RECOMBINANT
CHO CELL LINE AND HYBRIDOMAS WITH POROUS
MICROCARRIER CYTOPORE

CHENGZU XIAO, ZICAI HUANG, WENGQING LI, XIANWEN HU,
WENLU QU, LIHUA GAO AND GAOYAN LIU.

Department of Cell Engineering, Institute of Biotechnology,
20 Dongdajie, Fengtai, Beijing 10007l, P.R. China

Abstract Using porous microcarrier Cytopore and a low-serum medium supplement
BIGBEF-3, we have successfully cultivated recombinant CHO cell line CL-11G producing
prourokinase and hybridomas producing prourokinase monoclonal antibody in Celligen 1.5L or
5L bioreactor. The cell density could reached the yields of prourokinase and

monoclonal antibody increased with increasing cell density. As all of these cells could release

spotaneously and reattach to porous microcarriers, it was very easy to scale-up the cultivation.

Thus the bead to bead cells transfer method has been used to scale up the culture of CL-11G

cells in 20L bioreactor in the pilot production of prourokinase, and to scale-up the culture of
hybridomas in the production of McAB for purification of prourokinase.

Key Words Porous microcarrier; recombinant CHO cells; prourokinase; monoclonal
antibody

1. Introduction

The development of microcarrier culture by van Wezel in 1967 (1) made the large- scale
industrial culture of anchorage dependent cells for the production of vaccine and interferon
possible. In 1980s, a great advance of this procedure was achieved by the development of a
series of porous microcarriers (2-4). In previous paper we reported that using porous
microcarrier Cytopore, we had successfully cultivated a genetically-engineered CHO cell line
and some hybridomas to produce pro-UK and McAB (5). This paper reports our successful
results in using the bead to bead cells transfer method to sule up the cultivation of both the CHO
cells and hybridomas in the pilot production of pro-UK and McAB.

2. Materials and Methods
2.1 .CELL LINES AND MEDIA

The cell lines used were genetically-engineered CHO cell line CL-11G producing prourokinase
( 6 ) and hybridomas X15 and 38-1-7 cell lines producing prourokinase monoclonal antibody

(McAB) ( 7 ). The medium for CL-1G cells was DMEM:F12 (1:1), with the addition of some

amino acids and low-serum medium supplement BIGBEF-3 ( 5 ), and supplemented with 1%

NCS, aprotinin and kanamycin. The medium for

hybridomas was RPMI 1640, supplemented with 1% NCS and 0.1% pepton.

389

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 389-393.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

390

2.2. POROUS MICROCARRIER

The dry Cytopore microcarriers ( Pharmacia Co. Sweden ) were hydrated and swelled in PBS

before cultivation. After autoclaving at for 20 min, removed the supernatant and washed

the microcarriers with culture medium once or twice.

2.3. CULTURE VESSELS

Firstly spinner flask (Wheaton Co., USA) was used. The stirring rate was 30-40rpm. Then 1.5L
or 5L Celligen bioreactor (NBS Co., USA) supplemented with a modified perfusion controller (8)
and Biostat UC 20L bioreactor with spin filter was used.

2.4. CELLS COUNTING

The citric acid-crystal violet method was used to count cells in porous microcarriers. In order to
avoid a part of cells remaining in pores, the carriers were incubated for 4-5 hours and shaked
several times during incubation. At the same time, the MTT ( 9 ) method was also used to
confirm the satisfactory growth of cells inside the microcarriers.

2.5. GLUCOSE, PRO-UK AND McAB ASSAY

The methods used in measuring glucose and pro-UK were the same as that mentioned in
reference ( 10 ). The method for measuring McAB was ELISA assay ( 7 ).

3. Results

3.1. THE CULTURE OF CL-11G CELLS AND HYBRIDOMAS IN WHEATON SPINNER
FLASK

When the concentration of Cytopore was , and the medium was exchanged once or

391

twice each day, the cell density of both the CL-11G and hybridomas could reach over
( Fig. 1 ). The pro-UK and McAB yields increased with increasing cell density .

3.2. CELLS TRANSFER FROM BEADS TO BEADS

Just like the culture with Biosilon (10) and MC-1 (11) microcarriers, CL-11G cells could detach

from Cytopore porous microcarriers spotaneously and reattach on fresh ones. So it was also very

easy to scale up the cultivation as reported by Kamiya ( 12 ). When the scale up ratio was 3, as

shown in Fig 2,we withdrew 2/3 of culture from a spinner flask (200ml) and added the same

volume of fresh midium and microcarriers at 18th day, and then transfered the whole 200ml
culture into a larger flask with 500ml of fresh midium and microcarriers at 28th day, the final

cell density could still reach over each time (Fig.2, 3).

392

3.3. SCALE-UP CULTIVATION OF CL-11G CELLS FROM 5L BIOREACTOR TO 20L
BIOREACTOR

When we used the bead to bead cells transfer method to scale up the cultivation of CL-11G cells

from Wheaton spinner flask (700 ml ) to Celligen bioreactor ( 5L ), then to Biostat UC 20L

bioreactor, the cell density could also reach , and the highest yield of pro-

UK was ( Fig. 4 ).

3.4. SCALE-UP CULTIVATION OF HYBRIDOMAS FROM WHEATON TO 1.5L
BIOREACTOR

Using the same method we also successfully scaled up the cultivation of hybridomas from
Wheaton spinner flask ( 700 ml ) to Celligen 1.5L bioreactor. The cell density reached

, and the titer of McAB could reach ( Fig. 5 ). Using this McAB affinity
chromatograph, we purified the prourokinase to a purity of over 95% by HPLC assay with the
specified activity of (reported elsewhere).

4. Discussion

In a privious paper (5) we demonstrated that porous microcarriers had a lot of advantages
compared to solid microcarriers, such as : l.The concentration of carriers in the culture was
much less than that of solid microcarrier.2. Cells grown inside pores were protected for damage
by shearing stress, so that higher stirring speed could be used. 3. Cytopore porous microcarrier
was not only suitable to culture anchorage-dependent cells such as CHO cells, but
also suitable to culture anchorage-independent cells such as hybridomas. 4. As most of cells
were neted inside pores, cells suspended in the supernatant were much less. In addition, the
cells cultured inside the porous microcarriers were less depandent on attachment factor, so the
serum used in the culture might be furtherly reduced. This time we find an another advantage,

393

that is it can also be used to scale up the cultivation with bead to bead cells transfer method. So
we definitely consider that using porous microcarriers to large-scale culture genetically-
engineered cells for the production of recombinant products including prourokinase is a very
prospective approach.
5.References
1. van Wezel AL. Growth of cell-strains and primary cells on microcarriers in homogeneous

culture. Nature, 1967; 216:64
2. Looby D, et al. Immobilization of animal cells in porous carrier culture. TIBTECH, 1990;

8:204
3. Gotoh T, et al. A new type porous carrier and its application to culture of suspension cells.

Cytotechnology, 1993; 11:35
4. Shirokaze J, et al. High density culture using macroporous microcarrier. in: Spier RE.

Griffith B eds. Products for today, prospects for tomorrow. Oxford: Butterwerth-
Heinemann, 1994; 261
5. Xiao CZ, et al. High density cultivation of recombinant CHO cell line and two strains of
hybridomas with porous microcarrier Cytopore. Bull Acad Mil Med Sci, 1996; 20:191
6. Li FZ, et al. High-level expression of pro-urokinase cDNA in Chinese hamster overy cell
line (1). Bull Acad Mil Med Sci, 1993; 17:89
7. Ai X, et al. Preparation and characterization of monoclonal antibodies against uPA and
their application for purification of uPA. Bull Acad Mil Med Sci, 1995; 19:248
8. Xiao CZ, et al. An autocontroller system for cell perfusion cultivation on microcarrier.
Chinese Patent, ZL 91208893.1, 1994-09-14
9. Jia XH, Xiao CZ. MTT method for the esstimation of the number of cells cultivated in
microcapsules. Bull Acad Mil Med Sci, 1993; 17:207
10. Xiao CZ, et al. High density cultivation of a recombinant CD-1 cell line producing
prourokinase using a Biosilon microcarrier culture system. Chin Med Sci J,1994;9:203
11. Xiao CZ, et al. High density cultivation of genetically-engineered CHO cell lines with
microcarrier culture systems. Chin Med Sci J, 1994; 9:71
12. Kamiya K, et al. Subculture for large scale cell culture using macroporoous microcarrier.
in: Beuvery EC. et al. eds. Animal Cell Technology: Development towards the 21st
Century. Kluwer Acad Publ, The Netherlands, 1995;759

CELL-SETTLER PERFUSION SYSTEM FOR THE PRODUCTION AND
GLYCOSYLATION OF HUMAN INTERFERON- BY CLUMPED CELLS

D. LAMOTTE1, J. STRACZEK2 & A. MARC1
1 LSGC-CNRS, BP 451, 54001 Nancy Cedex - France
2 Lab. de Chimie - CHU, 54035 Nancy Cedex - France

1. Introduction
The production of recombinant proteins by mammalian cells is commonly performed in
batch or fed-batch cultures. Nevertheless, productivity in such a culture can be low. In
cell lines where protein synthesis is proportional to the cell division, a variety of means
have been designed to increase the maximum cell density. Amongst them are perfusion
systems, in which cells are retained whilst spent medium is replaced continuously with
fresh medium1,2.
Natural aggregation of animal cells has also been widely studied and proven to be an
efficient means of culture3 and sedimentation of aggregates in a column allows the

efficient retention of cells4. However, we report here the first animal cell perfusion

culture based upon cell aggregate sedimentation. This shape is very convenient for cells
sensitive to shear stress that can not be cultivated in classic perfusion systems.
Culture conditions may dramatically affect glycosylation and therefore the efficacy of
glycoproteins. Because the cell environment could change in perfusion cultures, the
glycosylation pattern of the protein of interest may be altered in a long-term culture, and
hence the product quality could be compromised. Few studies have reported the protein
glycosylation changes in perfusion cultures5 despite the emergence of cell perfusion
systems for industrial purposes6.
In this study, we have cultivated the CHO 320 cell line for the production of human
interferon- (IFN) in an 'in-house' perfusion system based on cell retention by
gravitational sedimentation. The IFN glycosylation was monitored by capillary
electrophoresis throughout the culture.

2. Materials & Methods
The producing IFN CHO cell line (CHO 320) was supplied by the Wellcome
Foundation Laboratories. Cultures were performed with a serum-free medium. We have
modified a 2L-stirred-tank bioreactor to allow the perfusion, and it was fitted with an
internal glassware tube (diameter: 1 cm, length: 12.5 cm) to provide a cell settling zone.
The bottom of the tube was lower than the agitation blades. The supernatant was
removed continuously from the vessel through the top of the glassware tube using a
peristaltic pump. The working volume in the reactor was 1.1L during the perfusion
phase. Prior to perfusion, cells were grown in batch culture for 80 hours. Separations
and detections of the IFN glycoforms were performed with a Beckman P/ACE 2100
capillary electrophoresis according to the protocol of James et al.7

395
O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 395-397.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

396
3. Results
3.1. KINETICS OF THE CHO 320 PERFUSION CULTURE

Cells were seeded in the reactor with an initial density of After being
propagated in batch culture for 80 hours to reach , the perfusion was
initiated in standard medium and the dilution rate increased daily from 0.022 to 0.055
until 350 hours of culture. During this phase, the IFN level increased rapidly to
reach the mean value of In parallel, initial clumps (100 m diameter)
enlarged with cell growth to reach a maximum diameter of 800 m after 550 hours of
culture (phase I). Thereafter, the perfusion rate was kept constant leading to the
stabilisation of the cell density at . The addition of amino-acids (excepted
glutamine) in the standard feeding medium (phase II) gave a transient decrease of the
IFN level. Conversely, the addition of vitamins in medium did not have significant
effect on the kinetic parameters (phase III).
Figure 1 shows that perfusion culture generated a twofold higher IFN accumulation
than batch culture. Consequently, the volumetric IFN production (in IU/1/hr) is
significantly higher in perfusion culture (x8). These results confirm that perfusion
systems are powerful tools to increase productivity of recombinant proteins.

397

3.2. IFN GLYCOSYLATION MACRO-HETEROGENEITY

The IFN glycosylation macro-heterogeneity was monitored during the perfusion culture
by use of capillary electrophoresis. The three classes of IFN glycoforms (doubly (2N),
singly (IN) and non glycosylated (ON)) were observed and their proportions reported in
figure 2. The proportion of fully glycosylated IFN decreased throughout the initial batch
phase of culture. Conversely, the glycoform proportions were constant during the four
phases of the perfusion culture and similar to those found at the end of the batch culture.
The cell-settler system described here allows the good product consistency for more
than 1000 hours despite changes in the culture environment and the kinetic parameters.

4. References

1. Goergen J.L., Lourenso da Silva A., Marc A. and Engasser, J.M. (1995) Fast propagation of hybndoma
cells in a forced-flow membrane perfusion reactor, in E.G. Beuvery, R.E. Spier and W.P. Zeistemaker (eds.).
Animal Cell Technology: Developments towards the 21st century, Kluwer Academic Press. Dordrecht, pp.
705-710.
2. Pinion H., Rabaud J-N., Engasser J.M. and Marc A. (1991) Cytoflow: a new perfusion bioreactor for
research and production, BFE 8, 344-347.
3 Chevalot I., Visvikis A., Nabet P., Engasser J.-M. and Marc A (1994) Production of a membrane-bound
protein, the human gamma-glutamyl transferase, by CHO cells cultivated on miccrocarriers, in aggregates and
in suspension, Cytotechnology 16 121-129.
4. Moreira J.L., Alves P.M., Rodrigues J.M., Cruz P.E., Aunins J.G. and Carrondo M.J.T. (1995) Growth of
BHK aggregates in a 2 liter bioreactor, in E.C. Beuvery, R.E. Spier and W.P. Zeistemaker (eds.). Animal Cell
Technology: Developments towards the 21st century, Kluwer academic press, Dordrecht, pp. 805-809
5. Gawlitzek M., Valley U., Nimtz M., Wagner R. and Conradt H.S. (1995) Characterization of changes in
the glycosylation pattern of recombinant proteins from BHK-21 cells due to different culture conditions, J.
Biotechnol. 42 117-131.
6. Bödecker B.G.D. (1994) Production of recombinant factor VIII from perfusion cultures: 1. large-scale
fermentation, in R.E. Spier, J.B. Griffiths and W. Benhold (eds.). Animal Cell Technology: Products of
today, prospects for tomorrow, Butterworths, London, pp. 580-583.
7 James D.C., Freedman R.B., Hoare M. and Jenkins N. (1994) High-resolution separation human interferon-
gamma glycoforms by micellar electrokinetic capillary chromatography, Anal. Biochem. 222 315-322.

DISPOSABLE BIOREACTOR FOR CELL CULTURE USING WAVE-INDUCED AGITATION

Vijay Singh

Schering-Plough Research Institute
1011 Morris Avenue
Union, NJ 07016 USA
email: [email protected]

Abstract

This work describes a novel bioreaclor system for the cultivation of animal, insect, or plant cells using
wave agitation induced by a rocking motion. This agitation system provides good nutrient distribution,
off-bottom suspension, and excellent oxygen transfer without damaging fluid shear or gas bubbles. Unlike
other cell culture systems, such as spinners, hollow-fiber, roller bottles, scale-up is simple, and systems
with up to 10 liter liquid volume have been operated successfully.

The device is disposable and therefore requires no cleaning or sterilization. Additions and sampling can
be done without the need for a laminar flow cabinet. The unit can be placed in an incubator requiring
minimal controls. All these features dramatically lower the purchase and operating cost of a lab scale
cell cultivation system.

Results are presented for various model systems: 1) recombinant NS0 cells in suspension; 2) human 293

cells in suspension/adenovirus production; 3) sf9 insect cell/baculovirus system; and 4) human 293 cells

on microcarrier. These examples show the general suitability of the system for suspension cells,
anchorage-dependent culture, and virus production. The system is especially well suited to virus and
vaccine production because of the high degree of containment.

Background

Technology for the cultivation of animal, plant and insect cells depends on the batch volume required. In
the laboratory, devices such as spinner flasks, roller bottles, T-flasks and similar systems are typically
used However, these devices can only produce 1 to 2 liters or culture per batch due to inherent oxygen
transfer limitations. In particular, spinner flasks are very popular for suspension culture as well for
anchorage-dependent systems on microcarriers, but they can only be used for a maximum liquid volume
of 1 liter.

There is no simple system to make 1 to 10 liters of cell culture which is the volume often needed for
protein characterization, inoculum production and pilot production. It is usually necessary to use
bioreactors which are typically stirred-tank bacterial fermentors modified to reduce shear forces. These
bioreactors are complex, expensive, and do not really provide an optimum environment for cell growth
due to high local shear and bubble aeration.

Adapting stirred tank technology to cell culture is a futile exercise because this design has intrinsically
high local shear rates. Instead it is critical to recognize the special demands of cell culture and design
directly to satisfy these needs. It is also essential to make the cultivation system as simple as possible to
operate. There have been many designs in the past, for example fluidized bed bioreactors and hollow-fiber
systems which have worked well, but were too complex to displace the spinner flask as the workhorse of
cell cultivation.

The objectives of this work was to develop a new cell culture bioreactor that:

1. Can be used to cultivate 100 ml to 10 liters of cells A high turndown ratio (maximum

volume : m i n i m u m volume) is also required to reduce the number of transfers. The poor 2:1 turndown

399

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 399-407.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

400

ratio of a conventional spinner flask makes inoculum scale-up very labor intensive, because the culture
must be transferred from one size spinner to the next larger size many times under sterile conditions.

2. Eliminates sparging of gas. It is clear that hubbies cause damage, so any good cell culture device should
not generate bubbles. Gas diffusion membranes are not satisfactory because of cost, complexity and poor
reliability. Also scale-up is limited by the membrane surface available for diffusion.

3. Eliminates mechanical agitation. Mixers generate high local shear. Seals around the mixer shaft are
costly and have poor reliability.

4. Guarantees sterility. The bioreactor should be available presterilized and disposable after use. This
eliminates the need for heat sterilization and cleaning. A large portion of the cost of a bioreactor is the
sterilization system. Cleaning of a bioreactor can take 8-10 hours per run making this a major operating
expense.

5. Containment. The unit should operate as a closed system so that it can be operated in the laboratory or
plant without the need for a laminar flow cabinet.

6. Simple controls. Like a spinner, the unit should not require complex controls. In cell culture, media can
be buffered to control pH and if the aeration system is capable of delivering the maximal oxygen demand
there is no need for DO control.

7. Low cost. A stainless steel bioreactor (10 liters) costs about $ 80,000. It should be possible to develop a
cell cultivator for 1/10 of this cost. The lack of moving parts, seals and instrumentation should also
significantly reduce operating expenses.

8. Easy to operate. Conventional bioreactors and pump-around systems require extensive training to
operate. Even small mistakes can lead to contamination. Just mastering the sterilization and calibration
procedures requires 1 -2 weeks of training. Many scientists just want to grow cells to study the protein
being expressed rather than become bioreactor experts. The cell cultivation system should therefore be
simple to learn. The goal was to develop a bioreactor which could be operated by an undergraduate
student with 1 hour of instruction.

A bioreactor system was developed that meets all these objectives. It is low cost, simple to operate, and
provides an optimal environment for cell and virus culture. This paper summarizes the oxygen transfer
capabilities, mixing time and cell cultivation potential of this technology based on wave-induced
agitation.

Materials and Methods

Bioreactor System and Operation

The bioreactor consists of a presterilized plastic bag that is partially filled with media and inoculated with
cells. The remainder of the bag is inflated with air which is also continuously passaged during the
cultivation. Mixing and mass transfer are achieved by rocking the bag back and forth. This rocking
motion generates waves at the liquid-air interface, greatly enhancing oxygen transfer. The wave motion
also promotes bulk m i x i n g and off-bottom suspension of cells and particles such as microcarriers.

The bag is discarded after harvest, eliminating any need for cleaning or sterilization since this is the only
component in contact with the cells. New bags are delivered guaranteed sterile by gamma radiation. Bags
are made of FDA approved biocompatible polyethylene. These bag materials are commonly used for blood
collection and biological fluid handling. Special ports have been developed to allow sterile additions to the
bag and to withdraw samples, without the need to place the bag inside a biosafety cabinet.

401

The rocking mechanism was developed by the author. The device as shown in Figure 1 consists of a
platform that rotates or rocks in one axis through an angle of 5 to 10°. Pneumatic bellows are used to rock
the platform at an adjustable rocking rate between 5 to 40 rocks/minute (rpm). A pre-sterilized bag is
placed on the platform; partially filled with media and then inflated using the sterile inlet gas filter
integral to the bag. Air is continuously passaged through the headspace of the bag by a pneumatically
driven pump. This provides oxygenation and gas exchange for pH control and CO2 removal. Exhaust air
passes through an exit sterilizing filter and a backpressure control valve. The unit requires only
compressed air (2 - 3 bar) to operate. No electrical connection is necessary. Temperature and pH control is
achieved by placing the entire unit inside a conventional cell culture CO2 incubator.

The rocking platform systems: CellbagTM, including specially designed cultivation bags (2 liter and 20 liter), are available from BioPro
International, Farmingdale, New York, USA ([email protected]).

Oxygen Transfer Measurement
The oxygen transfer capabilities of the system were assessed by measuring the volumetric oxygen transfer
coefficient kLa at various rocking rates, liquid volumes and aeration rates. The dynamic method was used
for kLa measurement. Here, the liquid in the bag is deoxygenated by passing nitrogen through the
headspace. Once the dissolved oxygen (DO) concentration was near zero, air was introduced in to the

headspace and the rise in DO recorded. kLa was calculated from the slope of the following mass balance
equation:

For the oxygen transfer studies an oxygen sensitive dye was added to enable non-invasive dissolved
oxygen measurement. Selected runs were confirmed using conventional DO probes inserted into the bag.
Mixing Time Measurement
Mixing time in the bag at various rocking rates were determined by injecting a fluorescent tracer dye into
the bag and videotaping the dispersion of the dye in the bag. UV light was used to enhance the contrast.

402

Time-tagged images were captured from the videotape and the mixing time was determined visually from
photographs. Mixing time was defined as the time required after injection to achieve complete
homogeneity.

Dissolved Oxygen and Carbon Dioxide

Off-line measurement of dissolved oxygen and dissolved were done by sampling the bag using a
syringe and analyzing it with a blood gas analyzer (Radiometer ABL5). Cell counts were done by
hemocytometer, and product assays were done by appropriate methods such as ELISA.

Results - Oxygen Transfer and Mixing

I. Oxygen Transfer Studies:

The rocking motion of liquid in the bags generates a larger mass transfer surface that static culture which
leads to much greater volumetric transfer coefficient Figures 2 and 3 show the achieved at various
rocking and aeration rates in 2 liter bags and 20 liter bags.

The oxygen transfer studies show that a of 3 hr-1 is possible in a 1000 ml liquid volume (2Lbag).
This is almost three-fold higher than the for 1000ml in a comparable 2L spinner (Table 1). This
should result in potentially 3X higher maximal cell densities in the bag compared to a spinner.

The data for spinner flasks in good agreement with published values. Aunins, et al (1989) report a
around 1 hr-1 for 500ml liquid in a 500ml spinner. Dorresteijn et al (1994) report a value around 2 hr-1 for
300ml liquid in a 600ml spinner system.

For the l0liter scale (in 20L bag) the maximum is around There is no comparable spinner

system since the surface to volume ratio decreases substantially as the spinner volume is increased. For
example, a 5 liter spinner would have a less than making it essentially useless for cell

cultivation.

With of 3 to and assuming a typical oxygen demand of (Singh, 1996), it is

possible to grow up to while maintaining dissolved oxygen concentrations above 10%

saturation.

With this knowledge of oxygen transfer it is possible to guarantee against oxygen depletion at any given
cell density by selecting the appropriate rocking/aeration rate. Unlike stirred tank bioreactors, there is no
bubble damage or danger of “overaeration”. stripping can be controlled by adjusting the
concentration of the inlet air. This effectively eliminates the need for a dissolved oxygen probe/ control

403

system. Samples may be taken and assayed periodically on a blood gas analyzer, off-line, to verify the DO
and levels.

II. Mixing Studies:

Estimates of mixing time were made by injecting a fluorescent dye and videotaping the dispersion of the
dye under various conditions. These experiments showed that the wave-induced motion was very effective
in mixing the liquid in the bag. Mixing time (defined as time for complete homogeneity) was typically 5
to 10 seconds. Subsequent studies with microcarrier cultivation showed good off-bottom suspension with
some particle gradients in the liquid. However, no significant settling of microcarriers or cells was
observed. The wave-agitation phenomena is the subject of a pending patent application.

Results - Biological Systems

1. Monoclonal Antibody Production - 1 liter and 10 liter scale

An important application for cell culture is the production of monoclonal antibodies in vitro. Spinner or
shake flasks can be used for 1 to 3 liters or culture, however oxygen transfer limitations in these system
preclude further scale-up. For larger volume it is necessary to use complex bioreactors based on stirred
tanks, hollow fiber, or immobilized technology.

The applicability of the Wave Bioreactor system for lab-scale (10 liter) culture was evaluated. The cells

used were a NS0 cell line expressing a humanized monoclonal antibody. Culture was performed in serum-

free media commonly used for this cell line. Cultivation was started in a 1 liter culture volume with 250ml
of inoculum generated in a 500ml spinner flask. After 96 hours the cell density was

cells/ml.

Two liters of fresh media was added to the bag and the cultivation continued. After an additional 96 hours

the cell density again reached and 7 liters of fresh media was added to the bag to bring to

the final culture volume of 10 liters, and cultivation was continued until zero viability (300 hours). Figure
4 shows the cell counts and monoclonal antibody production. The stepwise dilution technique used

demonstrates how the large turndown ratio (maximal volume/minimum volume) of the bag enables large

culture volumes without the need for transfers as would be the case using spinner flasks.
Culture profile was very consistent with smaller scale spinners. Cell densities exceeded

Dissolved oxygen levels remained above 50% saturation. Antibody expression was normal at over 600
mg/l. The system is clearly well suited for suspension cell cultivation and monoclonal production.
Inoculum scale-up can be done within the system without the need for splits.

404

2. 293 cells/Adenovirus Production

The potential of the Wave Bioreactor for virus production was tested using a recombinant adenovirus

system. A human embryonic kidney cell line (293) grown in suspension was used as the host. Initial

seeding density in a 2L bag containing 1000ml of serum-free media was The

unit was placed in a 37°C incubator with a 5% CO2 atmosphere. The culture was grown for 5 days at a
rocking rate of 10 rpm and 0.1 vvm aeration. At this stage, the viable cell count was
500 ml of the culture was removed and replaced with 500 ml of fresh media. After an additional 24 hours

the culture was infected with an MOI of 30 virus particles/cell. The culture was harvested three days post-

infection at which point the viable count was essentially zero (Figure 5). The final virus titer was 10,000
particles/cell.

From Figure 5, it is apparent that the system is quite capable of maintaining adequate dissolved oxygen
levels at these cell densities. The pH remains very constant due to the excellent gas exchange as evidenced
by the constant until late in the infection when the pH drops rapidly beyond the capability of the
buffer system.

The system is very easy to use. All additions and sampling were done in the incubator itself. No biosafety

hood was required. The standard 0.1 micron exhaust filter was supplemented by a Pall DFA cartridge
filter. The integrity of the exhaust filter system was tested by passing the exhaust gases to an uninfected
293 culture. Absence of any infection in the second culture confirmed the effectiveness of the exhaust
filtration in containing adenovirus.

3. Baculovirus Production - 10 liter scale

The sF9/Baculovirus system is an excellent method for the rapid expression of large amounts of
recombinant proteins. However, the oxygen requirements of sF9 cells are higher than mammalian cells
and this restricts the volume of spinner or flask culture to around 1 liter. Typically, larger volumes than
are possible in flasks are required to produce sufficient protein for isolation and complex bioreactors are
usually necessary. It was desired to determined if the simple Wave Bioreactor system could be used to
scale-up inoculum and produce virus in 10 liter liquid volume.

The gene sequence for a recombinant protein was cloned into the pAcHLT-B baculovirus expression
vector, transfected into sf9 cells and a high titer virus slock was generated.

For cultivation in the bag, 1 liter of Bio-Whittaker X-Press Insect Cell Media was added to a 20 liter (total

volume) bag. This was inoculated with Rocking rate was 20 rpm in a humidified

405

incubator at 25 °C. After 6 days of cultivation the cell count reached At this point 4

liters of fresh media was added to the bag. Within an additional 4 days the cell count reached
in this increased volume. At this point another 4 liters of fresh media was added along with the
virus at a MOI of 0.5 bringing the system to a total volume of close to 10 liters. Harvest was done 2 days

post infection.

This system produced a titer comparable to 100 ml shake flasks. It demonstrated the ability to increase cell
mass by simply added fresh media to the system. No conventional series of flasks or splits were required,
reducing labor and potential for contamination. All additions and transfers into the bag were done in the
incubator. Figure 6 shows that oxygen transfer and CO2 desorption was not limiting.

4. Microcarrier Culture

Human 293 cells were cultivated on Cytodex 3 microcarriers. Growth media was DMEM with 10%
bovine calf serum. Inoculum was collected by trypsinizing microcarrier cultures grown in spinner flasks.
60ml of inoculum was added to 3g Cytodex 3 in 900 ml of media and then transferred to a 2 liter bag. 10
ml of the inoculum was added to a spinner flask with 0.45g Cytodex 3 and 140ml media (250ml Bellco)
that was run in parallel at 40rpm. Growth in the bag culture was very similar to the spinner. Cell counts,
glucose consumption and lactate formation were equivalent with the exception of slightly higher cell
counts in the supernatant from the bag. Cells grew to confluency ( >50 cells/bead) in the bag .

Applications

1. Substitute for Spinner Flasks and Roller Bottles

Spinner flasks cannot be used for liquid volumes greater than 1 liter because of insufficient oxygen
transfer. The Wave Bioreactor generates much more mass transfer surface because of the unique wave
action. Operation to 10 liters is routine with cell densities over

406

The Wave Bioreactor has a greater range of operating volumes. This reduces the number of transfers that
need to be made minimizing handling and potential for contamination.

The Wave Bioreactor is an ideal way to make a few liters of cell culture for lab purification, inoculum
propagation and cell bank generation.
2. Ultra Low cost Bioreactor:
Traditionally the only way to produce 5 to 20 liters of cell or virus culture was to use expensive and
complex bioreactors. The Wave Bioreactor is 1/10 the price of a comparable bioreactor and does not
require substantial training. No utilities such as steam, cooling water and process gasses are required. No
complex instrumentation is necessary. The unit can be set up in a conventional cell culture incubator. A
CO2 controlled incubator can be used if pH control is desired. This makes lab scale cell and virus
cultivation possible even in small labs, hospitals and universities.

3. High Containment / GMP Systems:
The u n i t is completely closed. Additions and samples can be taken using a hospital-type needleless syringe
connector or tube-fuser. These operations do not require a laminar flow cabinet but can be done with the
unit inside the incubator. The bag is certified to be sterile by gamma radiation. Air into and out of the bag
is sterilized by filters appropriate to the containment level necessary. At the end of cultivation the bag can
be removed to another area for processing. The disposable nature of the system makes is very suitable for
virus production and BL2/BL3 operations.

407

4. Cell Therapy and Gene Screening:

The Wave Bioreactor allows even the smallest lab or clinic to perform cell culture. The simplicity of the
systems enables the scientist to concentrate of the use of the cells rather than learning how to operate
complex bioreactor systems. The large cultivation volumes and the ability to grow attachment-dependent
lines make it useful for primary cell cultivation for cell therapy applications and the simplicity of the
system allows the rapid expression and screening of hundreds of genes needed for gene therapy and
human genome elucidation.

Conclusions

A novel bioreactor (patent pending) for cell culture has been developed. The system utilizes an inflated

disposable plastic bag as the cell cultivation chamber. These specially designed bags are made of

biocompatible materials and are delivered sterile, eliminating the need for cleaning and sterilization.
Oxygen transfer and mixing are accomplished by wave-induced agitation caused by rocking the bag back

and forth. The rocking mechanism has been optimized in terms of rocking angle, rocking rate, aeration,

and mechanical design to provide a kLa for oxygen transfer of 2 to Based on typical cell respiration

rates, this transfer rate is sufficient to grow up to The bioreactor system, unlike spinner
flasks, is not limited by gas-liquid transfer surface and scale-up to 10 liters operating volume has been
demonstrated using several typical cell and virus production systems.

This simple, low cost wave bioreactor system can be used for animal, insect and plant cell culture. The
closed design is very suitable for virus production or other applications requiring high containment. The
system requires minimal instrumentation and can be operated inside a laboratory incubator. All handling
can be done in the open, eliminating the need for a laminar flow hood.

References

Aunins. J.G., Woodson, B.A.. Hale, T.K. and Wang, D.I.C.. 1989. Effects of Paddle Impeller Geometry on power Input and Mass Transfer
in Small-Scale Animal Cell Culture Vessels. Biotechnol. Bioeng. 34: 1127-1132.
Dorresteijn, R.C., de Gooijer, C.D., Tramper, J. and Beuvery, E.C., 1994. A Simple Dynamic Method for On-line Determination of k L a
During Cultivation of Animal Cells. Biotechnol. Techniques. 8, 9 675-680.
Singh.V., 1996. On-Line Measurement of Oxygen Uptake in Cell Culture Using the Dynamic Method. Biotechnol. Bioeng. 52: 443-448.

Acknowledgments:

*US and European patents are pending for portions of this work. I would like thank my many
collaborators especially Nancy Connelly, John McDowall, Edward Glowacki, Raphael Nir, Colleen Brolly,
Peter Pappas and Jeffery Mutter.

USE OF miniPERM SYSTEM FOR AN EFFICIENT PRODUCTION OF MOUSE
MONOCLONAL ANTIBODIES

M.L.NOLLI*, R.ROSSI°, A.SOFFIENTINI°, D.ZANETTE°, C.QUARTA°
°Biosearch Italia, Via R Lepetit 34, 21040 Gerenzano (VA) Italy
*Consultant Biotechnology, Biosearch Italia

ABSTRACT
A rapid and efficient method for the production of monoclonal antibodies has been set up
using the miniPERM system. The miniPERM represents a new generation of compact and
disposable laboratory bioreactors for the growth of suspension cells. The combination of
some important requirements as 1) high surface volume ratio; 2) sterility; 3) small space
occupied; 4) monouse; 5) easy to use; 6) possibility to use four modules at the same time
makes the system very useful for an efficient production of hundreds of milligram quantities
of MAbs. A mouse hybridoma cell line, from NSO Ag 14 non secreting cells, has been
cultivated in miniPERM modules and the corresponding IgG1 MAb was collected after
1 week run. 70-100 mg of MAb/module/week have been produced either in conventional
medium (with serum) or in serum free medium. This module productivity permits a
production of more than 300 mg of MAb/week using at the same time the four modules.
The SDS-PAGE and IEF analysis of the purified IgG1 MAb is presented.

1. INTRODUCTION
At early phases of a biotechnology project, aimed to produce either monoclonal antibodies
(MAbs) or recombinant proteins, it is crucial to have available several mg quantities of the
protein, possibly in one batch, to set up purification and analytical methods and to carry out
preformulation studies.
Small disposable bioreactors, specifically designed for high density cell cultures, offer the
possibility to easily cultivate either hybridomas or recombinant cells using a high-surface
volume ratio concept and produce the needed secreted product. Furthermore, the
combination of the use of these flexible bioreactors with appropriate serum free media
permits reliable cell cultivation and product generation (1).
We describe here the production of a mouse monoclonal antibody from NSO myeloma
parent cells in miniPerm bioreactor (2). The production was shown to be very efficient. 70-
100 mg/module/week of MAb were produced in serum free medium. It was also
demonstrated that, when using simultaneously four modules, there is a consistency in the
MAb production in terms of number and composition of isoforms and quantity of MAb
produced.

409
O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 409-415.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

410

2. METHODS
2.1 CELLS

The mouse hybridoma, deriving from NSO parent myeloma cells, was adapted to grow in

serum free medium
(SFM,Gibco cat. N: 12045-084) using the following protocol:
2.1.1. Cells were plated at in RPMI + 10% FCS + 5% Origen (Igen).
2.1.2. Then they were progressively diluted with SFM, 1:2, 1:4, 1:16, 1:32 and then in
SFM alone. Every dilution with SFM, cells were left to grow for several generations,
monitoring the cell viability and MAb production.

2.2 BIOREACTOR

miniPERM bioreactor (Heraeus Instruments GmbH) is a small size bioreactor made up of
two connected modules:
2.2.1. a production module (disposable culture chamber)
2.2.2. a nutrient module (reusable culture chamber)
The bioreactor is rolled on a special turning device, specifically designed for a 37 °C
incubator. This roller apparatus permits the simultaneous use of four modules.

2.3 MODE OF CULTURE (FED-BATCH)

2.3.1 .The two miniPERM modules (nutrient and production module) were assembled in
sterility following the operating instructions.
2.3.2. , at the end of the log phase, in 35 ml of SFM were then inoculated
into the production module.
2.3.3. 350-400 ml of SFM were used to fill the nutrient module.
2.3.4. The apparatus was put onto the roller apparatus and then rolled.
2.3.5. 7-8 days runs were carried out, changing medium twice and taking samples every
day.

2.4 .MAb TITRATION

The MAb produced was tested by an antibody capture ELISA set up in house.
The antibody quantitation was done by using a standard curve of the purified MAb. The
results were calculated with the use of logit-log regression.

2.5. MAb PURIFICATION AND ANALYSIS

2.5.1. The MAbs in the ascitic fluids and miniPERM supematants were purified using
Protein G Sepharose 4 FF (Pharmacia cat N° 17-0618-02). The samples were loaded onto
the column with the conductivity and pH of the equilibration buffer (0.02M Na-phosphate
pH=7.0). The column was washed with the same buffer. The MAbs were eluted with 0.1M

411
Gly-HCl pH=2.7 and immediately neutralized with 1 .0M Tris-HCl pH=9.0.
2.5.2. SDS PAGE (reduced): The reduced SDS-PAGE was done with a precast NOVEX™
Tris-Glycine gel with a 4-20 % gradient (Cat. N° EC6025). -mercaptoethanol was used
as reducing agent. The running time was about 90 min. at a constant voltage of 120 V. At
the end of the run the gel was fixed and then stained with Coomassie blue.
2.5.3. Isoelectric Focusing (IEF)(3): The IEF analysis was done using precast NOVEX™
IEF pH 3-7 gels (Cat. N° EC6655B).
After running, the gels were fixed and then stained with Coomassie blue.
2.5.4.Densitometric measurements: The densitometric measurements were done with the
SHIMADZU CS-9301PC dual wavelength flying spot scanning densitometer.

3. RESULTS AND DISCUSSION
The miniPERM is a small modular bioreactor for the production of MAbs for analytical
use.
The results of the MAb production using this system are reported in Table 1. Four modules
were simoultaneously used on the same roller apparatus to carry out one-week runs.
The fed-batch mode of culture was used, changing medium twice/week.
As shown, the quantity of MAb produced and its concentration in the medium, is
comparable in the four modules. In fact the total quantity of MAb is in a range of 80-100
mg/module, while the concentration of the MAb is in the range of 2.5-3 mg/ml in 32 ml
volume /module.
The MAb production is very efficient. The recovery of MAb in the medium is at a high
concentration (2.5-3 mg/ml) and in a limited volume (32 ml).
The quantity of MAb, 80-100 mg /module /week, is an important amount for analytical
use. In addition, the simultaneous use of the 4 modules makes possible a production of
more than 350 mg of MAb in one week. An additional advantage is the production in
serum free medium that allows either the use of MAb as it is or for an easier purification.
Once purified, the MAb produced in the four modules (A,B, C, D) was analysed by SDS-
PAGE and by isoelectric focusing along with the densitometric distribution of the
glycoforms. The results of these analyses demonstrate: 1) in reduced SDS-PAGE this MAb
contains the typical IgGl heavy and light chain bands as those produced by the MAb from
the ascitic fluid (fig. 1) and 2) both the IEF and densitometric analyses show the consistency
of the glycoform distribution of the MAb produced in the four modules (fig 2 and fig. 3).
Isoelectric focusing was also used to compare the respective glycoform distribution of the
MAb produced by miniPERM and ascitic fluid(fig.4). The densitometric analysis (fig. 5) of
the IEF gels show that the MAbs have the same number of glycoforms and similar
percentage distribution for every glycoform.

412

413

414

415
4. CONCLUSION
The consistency of the glycoform distribution and the amount of the MAb produced in the
miniPERM modules prove the reliability ofthe system for production of hundred mg
quantities of Mab in a single batch.

REFERENCES

1) Merck & CO., INC. Large scale production of mammalian cell infective viruses e.g. hepatitis A- in order to
produce a commercial vaccine. Pubblication: WO95/04812.
16 February 1995.
2) Falkenberg, F.W., Hengelage, T., Krane, M., Bartels, I., Albrecht, A., Holtmeier, N., and Wuthrich, M.: A
simple and inexpensive high density dialysis tubing cell culture system for the in vitro production of monoclonal
antibodies in high concentration.,
J. Immunol. Methods 165 (1993), 193-206.
3) Williamson, A.R.: Isoelectric focusing of immunoglobulins. In " Handbook of Experimental Immunology 1,
Blackwell scientific Publication Oxford, 1978.

PRODUCTION OF ANTI-EGF RECEPTOR hMAB IN HOLLOW
FIBER BIOREACTORS FOR IN VIVO DIAGNOSIS.

Arias, M.A., Valdés A., Curbelo D., Morejón O.M., Caballero I., Villán
J., Gómez J.A., Fernandéz A., Rodríguez J., Morales A., Boggiano T.,
Castillo A., Bouzó L., Hermida C., García G., Vitón P., Pérez N.,
Rodríguez T.
Center of Molecular Immunology, P.O.Box 16040, Havana 11600,Cuba.

1. ABSTRACT
NSO transfected myeloma producing an anti-EGF receptor humanized MAb was
cultivated in hollow fiber bioreactors up to production scale (Acusyst-P3/X). Its
production rate and metabolic uptakes were compared with a hybridoma, that secreted
the murine variant of this product also produced in hollow fiber. The control strategy of
the fermentation is discussed. The purification scheme, mainly based on affinity and ion
exchange, was established to obtain a product with a purity higher than 96 %. The freeze
dring process was studied to obtain a final reduced product for labeling with for “in
vivo” diagnosis.
2. INTRODUCTION
The production of an anti-EGF receptor humanized MAb was carried out in the GMP
facilities of the Center of Molecular Immunology, which is a Cuban Biotechnology
institution devoted to production of proteins from animal cell culture(1). NSO myeloma
was transfected to secrete a “reshaped” monoclonal antibody, which recognizes the
EGF-receptor (2). It has been demonstrated that this MAb is a very efficient as marker
of tumors with high level of anti-EGF receptor (lung, breast, brain, head and neck) (3).
3. RESULTS AND DISCUSSION

417
O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 417-419.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

418
3.1 Upstream process
The production process is summarized in Figure. 1. Hollow fiber bioreactors were used
to cultivate NSO transfected myeloma up to production scale. The inoculum preparation
was done in T flasks and spinners (feed-batch mode) to obtain
with viability higher than 90 %, using DMEM/F12 medium plus 5% of FBS.
During stationary culture it was demonstrated that the protein production rate increases
when the growth rate decreases due to nutrient limitation (4). The protein production
rate (PPR) at large scale was stabilize around 280 mg/d (Figure 2). The metabolic
profiles don't suggest any limitation in perfusion (Figure 3) and the dissolved oxygen
was kept high enough (data not showed). The PPR could be increased by manipulation
of the cell line.

The control strategy for nutrients was designed from kinetic studies in stationary culture
(5). It was shown that during the growth phase the glucose concentration had a minimum
range between 6 - 7 mM and the lactate concentration has a maximum of 12 mM.
Considering the difference between stationary culture and hollow fiber systems, the
main parameters to be controlled during the run were defined as follows. During the
growth phase, pH should be 7.2 ; minimum glucose concentration, 8.3 mM and
maximum lactate concentration, 11.1 mM. During production phase, pH should be 7.0,
minimum glucose concentration, 5.5 mM and maximum lactate concentration, 16.6
mM. The minimum partial pressure of oxygen should be 100 mm Hg during both
phases.
3.2 hMAb Purification Process
As it is shown in Figure 4, during anion exchange there are losses of about 20 % of the
initial mass, probably due to pH (7,3), which should be 1 unit lower than the isoelectric
point of the protein. According to DNA contents of the product, the relation A280/A260
was kept above 1,5 in the different streams after Protein A, implying a low level.

419

3.3 Formulation of freeze-dried product

To enhance the performance and quality during in vivo diagnosis a freeze-dried
formulation is used, which contains the reduced MAb ready for labeling with The
reduction process is done by the Schwartz method, Mather modified (3). Then the
product is desalted to nitrogen purged PBS. After that, a reducer agent is added.
The filled vials are lyophilized to with a first freezing to -45°C, a primary

drying to -20°C and a secondary drying to 25°C. The final product is submitted to

Quality Control criteria, which is summarized in Table 1.

4. CONCLUSIONS

The process showed produces a freeze-dried formulation of anti-EGF receptor hMAb
with the required quality to be used as an “in vivo” diagnosis drug, in the programmed
phase I clinical trial. The production process of this hMAb can be enhanced, mainly
during the upstream and purification steps.

5. REFERENCES

1. Villán J. Et. Al., Strategy for Scale Up of Mammalian Cell Culture in the New GMP Manufacturing
Facility of the Cuban Biotechnology Industry. In : Animal Cell Technology. From Vaccines to Genetic
Medicine. Proceedings 14th ESACT Meeting.

2. Mateo C., Moreno E., Amour K., Lombardero J., Harris W., Pérez R., (1997). Reshaped of humanized
monoclonal antibody to EGF-R with retention of full binding activity. Immunotechnology 3: 71-81
3. Iznaga N., Morales A., Ducongé J., Caballero I., Fernández E., Gómez J.A., (1998) Pharmacokinetics,
biodistribution and dosimetry of anti-human epidermal growth factor receptor humanized
monoclonal antibody R3 in rats. Nuclear Medicine & Biology, Vol. 25. (in press).
4. Boggiano T. (oral communication)
5. Castillo A. J., Fernández A., Boggiano T., (1997), Study of different cell culture conditions for the
production of “reshaped” Mab in NSO cells, ESACT Meeting poster.

OPTIMISATION OF THE PRODUCTION OF VLP'S IN 2 LT STIRRED

BIOREACTORS

P. E. Cruz1, A. Cunha1, C. C. Peixoto1, J. Clemente1, J. L. Moreira1,
M.J.T. Carrondo1,2
1 - IBET/ITQB, Apartado 12, P-2780 Oeiras, Portugal

2 - Laboratório de Engenharia Bioquímica, F. C. T./ U. N. L, P-2825
Monte da Caparica, Portugal

1. Abstract

In this work the production of virus like particles (VLP’s) in insect cells was optimised
for 2 Lt bioreactors. For this purpose the effects of and hydrodynamic conditions on
product quality were evaluated. Although the did not show a significant effect upon
cell growth in the range from 10 to 50%, cell infection and specific productivity were
dramatically affected. The production was optimal at a of 25% and decreased up to
two fold when the decreases to 10 or increases to 50%. The highest VLP titre was
obtained at 96 hours post infection (hpi) and at of 25%. The effect of overcritical
conditions upon productivity was also studied: cell infection is affected by agitation
rates above 270 rpm or by aeration rates higher than 0.04 vvm, with a three fold
decrease in VLP titre.

2. Introduction

Several strategies have been used to design a safe and efficient vaccine for AIDS
including whole inactivated virus (Murphey-Corb et al., 1989), recombinant HIV
subunits like gp120 (el-Amad et al., 1995) and genetically engineered virus like
particles (Rovinsky et al., 1992).

In this work, the insect cell (Spodopter a frugiperda, Sf-9) / baculovirus (AcNPV)
system expressing Pr55gag particles was used. The Pr55gag is targeted to the plasma
membrane and assembles in 100-120nm particles and therefore this product is
extracellular. In order to gel high quality product it is essential to ensure that particle
formation is performed correctly, since it has been shown that the titre of antibodies
after immunisation with virus like particles is significantly higher than after
immunisation with cell lysate or recombinant proteins (Vzorov et al., 1991). The cells
were infected at several values and product titre and quality were determined for the
hydrodynamic conditions previously optimised for this system (270 rpm and 0.04 vvm).

421

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 421-427.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

422

Finally, to confirm that these hydrodynamic conditions cannot be overcome, two
experiments were performed: one at higher agitation rate and the other at higher aeration
rate; the Pr55gag titre obtained was compared with the ones obtained in the optimal
conditions.

3. Materials and Methods

3.1 CELL GROWTH AND INFECTION EXPERIMENTS

Spodoptera frugiperda Sf-9 insect cells (ECACC no. 89070101), previously adapted to

SF900II (Gibco, Glasgow, UK) (Cruz et al, 1997) were maintained in
Wheaton stirred vessels (Wheaton, Millville, USA). Cell growth and infection

experiments were performed in a fully controlled stirred bioreactor (B. Braun,

Melsungen, Germany) inoculated at

Cell concentration was determined as previously described (Cruz et al, 1997).
Cells were infected with baculovirus that generate Pr55gag particles, provided by

Prof. Polly Roy (IVEM, Oxford, UK), in the mid exponential phase of cell growth with
a multiplicity of infection (MOI) of 2 pfu/cell. Samples were taken at several times post

infection and the supernatant was collected for VLP quality analysis.

The effect of upon cell growth and infection was studied at 10, 25 and 50%

controlled by the aeration rate up to a maximum of 0.04 vvm. The agitation rate in

these experiments was maintained at 270 rpm. To test the effect of overcritical

conditions in VLP titre two experiments were conducted: one to test the effect of an

overcritical aeration rale (0.06 vvm) and another to test the effect of an overcritical

agitation rate (300 rpm); the was maintained at 25% in both cases.

3.2 PR55GAG CONCENTRATION MEASUREMENTS AND QUALITY ANALYSIS

Pr55gag was quantified by using the Innotest HIV Antigen Mab (Innogenetics N.V.,
Zwijndrecht, Belgium), kindly provided by Innogenetics N.V. (Ghent, Belgium). Since

this test uses monoclonal antibodies against p24, related polypeptides containing the
recognised p24 epitopes will also be detected (e.g. breakdown products of Pr55gag).

In order to determine the fraction of Pr55gag in the form of high molecular

weight particles the culture supernatant was injected into a gel filtration column

(Pharmacia Biotech, Uppsala, Sweden) connected to a FPLC system (Pharmacia
Biotech). Phosphate buffer (10 mM pH7.2) was used as running buffer at a flow rate of

0.4 ml/min. The column was calibrated with high and low molecular weight standards
(Bio-Rad, Hercules, USA) and the void volume was determined by using blue dextran
2000 (Bio-Rad). Fractions of 1 ml were collected for 1.5 column volumes and the

activity was measured using the ELISA test described above. The fraction of product in
the form of particles was determined by calculating the ratio between the activity in the

high molecular weight fractions (0.9 to 2.4 MD) and the tolal measured activity.

423

4. Results and Discussion

4.1 INFLUENCE OF IN CELL GROWTH AND INFECTION

Figure 1 shows the results of three cell growth and infection experiments performed at

different values: 10, 25 and 50%. In all cases the agitation rate was maintained at

270 rpm and the was controlled by the sparged air flow up to a maximum of

0.04 vvm (the critical values obtained in previous experiments)(data not shown).

As can be observed in Figure 1A, maximum sustainable cell concentration (obtained at
270 rpm, 0.04 vvm sparged air) increased from to as the

was lowered from 50 to 10% resulting from an increase in the oxygen supply. As
can also be observed, the specific growth rate was the same in all the experiments (0.031

) which indicates that cell growth is not significantly affected by in this range. In
fact, the optimal values reported for cell growth range from 5 to 100, depending on
bioreactor size and configuration, culture medium and inlet gas (Schmid, 1996).

The optimal dissolved oxygen tension is known to be different for different
proteins (Agathos, 1996); to study the influence of upon cell infection, the variation
of normalised cell concentration, (the ratio between the cell concentration at a given
time and the cell concentration at infection), with time was obtained for 10, 25 and 50%

(Figure IB). As can be observed, cell death caused by virus infection is slightly
faster for 25% dissolved oxygen than at 10 or 50%. Since the decrease in viability is
related with virus infection, higher titres can be expected for the experiment with 25%

From these results it can also be concluded that a MOI of 2 is sufficient for
synchronous infection, since the cell concentration did not increase after infection. In
addition, since the infection is performed in the mid exponential growth phase, higher
MOIs would not give significant enhancement of the final product titre (Neutra et al.,
1992).

424 IN PRODUCT TITRE AND QUALITY
4.2 INFLUENCE OF

In order to determine if these differences in infection profiles affect the product and
particle titre, ELISA and gel filtration chromatography tests were performed with the
samples obtained from each experiment (Figure 2).

In terms of total product (Figure 2A) a much higher titre was obtained when was
maintained at 25%; higher or lower values decreased the maximum titre to half of
the one obtained at the optimal Moreover, the specific productivity at 25% was
similar to the one found in small spinner flasks under no oxygen limitation,

indicating that the cells are not affected by the increase in scale (data not shown). The
product quality determined by gel filtration chromatography is defined as the percentage

of product in the form of high molecular weight particles; this analysis is depicted in
figure 2B for 10, 25 and 50% By applying these percentages to the total titre
(Figure 2A) the particle titre could be obtained (Figure 2C). As can be observed from

Figure 2C, also the particle titre was larger at 25% than at 10 or 50%. By comparing

the results obtained for 10% and 50% it can be observed that, although the total titre

425
was similar in both cases (Figure 2A), the particle titre at 10% was almost twice the one
obtained at 50% due to the higher quality achieved at 10%

From Figure 2B, it is possible to conclude that the maximum particle titre is not
coincident with the maximum quality, i.e. with the maximum particle percentage. While
the maximum particle titre was obtained at 96 hpi (Figure 2C) the best quality was
obtained at 48 hpi, independently of the value (Figure 2B). In fact, it is at 48 hpi
that the cell lysis caused by baculovirus begins (King and Possee, 1992), indicating that
there may be some proteolythic degradation starting at that time, resulting in a decrease
in the relative amount of particulate material. From these results it is possible to
conclude that the harvest time has to be carefully optimised for each product and for
each bioreactor size.

In order to study the extent of the influence of conditions slightly more harsh than
the ones at which cells are affected, but less stringent than the ones at which cells are
killed two experiments were conducted. In the first experiment the agitation rale was
increased to 300 rpm and in the second case the aeration rate was increased to 0.06 vvm.
In order to maintain the a the aeration rate had to be decreased to 0.03 vvm in the first
experiment and the agitation rate had to be decreased to 200 rpm in the second
experiment. Figure 3 represents the particle titre obtained in these two experiments,
compared with the results obtained in the optimal conditions.

As can be observed, the small increase in the shear stress caused a dramatic decrease in
the case of higher agitation rate (74%) and a similar decrease in the case of higher
aeration rate (77%), when compared with the titre obtained in the previously determined
optimal conditions. Since in both cases of overcritical conditions neither a change in
morphology due to stress nor an abnormal decrease in cell viability were observed it can
be concluded that although the stress is below the cell killing threshold it is already
affecting productivity.

426

5. Conclusions

From this work the following conclusions can be drawn:
A) The did not show a significant effect upon cell growth in the range from
10% to 50%. Conversely, cell infection and specific productivity were dramatically
affected up to a two fold decrease when the decreases to 10 or increase to 50%, the
optimal value being 25%.
B) The maximum product quality is not coincident with maximum product titre.
Although the best quality product was obtained at 48 hpi, independently on the , the
highest particle titre was obtained at 96 hpi and at of 25% (according to the gel
filtration chromatography and Western immunoblot results).
C) It was demonstrated that cell infection is affected by agitation rates above 270
rpm and by aeration rates higher than 0.04 vvm, even when the overcritical values are
still far from the limit at which cell death starts to occur.
These results emphasise the importance of working within the limits where the
cells are not affected by the environmental conditions, not only in terms of but also
in terms of agitation and aeration rate.

ACKNOWLEDGMENTS

The authors acknowledge and appreciate the technical support of Ms. Maria do Rosário Clemente, the
financial support received from the European Commission (VALUE CE-CTT-535) and Fundação para a
Ciência e Tecnologia - Portugal (EUREKA PUEM/C/EU and BD/4545/94) and the partnership with
Innogenetics S.A..

REFERENCES

Agathos, S N. (1996) Insect cell bioreactors. Cytotechnol., 20, 173-189.
Cruz, P. E., Moreira, J. L., Carrondo, M. J. T. (1997). Insect cell growth evaluation during serum-free

adaptation in stirred suspension cultures. Biotechnol. Techniques 11, 117-120.
el-Amad, Z., Murthy, K. K., Higgins, K., Cobb, E. K., Haigwood, N. L., Levy, J. A., Steimer, K. S. (1995)

Resistance of chimpanzees immunized with recombinant gp120SF2 to challenge by H1V-ISF2. AIDS 9,
1313-1322.
Murphey-Corb, M., Martin, L. N., Davison-Fairburn, B., Monelaro, R. C., Miller, M., West, M., Ohkawa, S.,
Baskin, G. B., Zhang, J. Y., Putney, S. D., Allison, A. C., Eppstein, D. A. (1989) A formalin-inactivated
whole SIV vaccine confers protection in macaques. Science 246, 1293-1297.
Neutra, R., Levi, B.-Z., Shoam, Y. (1992) Optimisation of protein production by the baculovirus expression
vector system in shake flasks Appl. Microbiol. Biotechnol. 37, 74-78.
Rovinsky, B., Haynes, J. R., Cao, S. X., James, O., Sia, C., Zolla-Pazner, S., Matthews, T. J., Klein, M H.
(1992) Expression and characterization of genetically engineered human immunodeficiency virus-like
particles containing modified envelope glycoproteins: implications for development of a cross-protective
AIDS vaccine. J. Virol 66, 4003-4012.
Schmid, G. (1996) Insect cell cultivation: growth and kinetics. Cytotechnol. 22, 43-56.
Vzorov, A. N., Bukrinsky, M. I., Grigoriev, V. B., Tentsov, Y., Bukrinskaya, A. G. (1991) Highly
immunogenic human immunodeficiency virus-like particles are produced by recombinant vaccinia virus-
infected cells. AIDS Res. Hum. Retroviruses 7, 29-36.

Discussion 427
Anon:
What evidence do you have for having capsid particles, other than
Cruz: size occlusion? We have had trouble finding capsid particles that
Bernard: are stable when they are non-enveloped from Maloney Leukaemia
virus. You could just have a conglomeration of gag proteins. Do
Cruz: you have EM confirmation?

Yes, we have confirmed the presence of particles. However, the
particles are not that stable as after 1 month at 4° C there is a
considerable reduction.

I am unsure of your strategy for process control after you had set
maximal operating parameters - do you have constant agitation and
air gassing and on top oxygen supply, or do you vary to these
points?
These are limiting points. So we have done experiments both
maintaining agitation and aeration rates, whilst the other variables
went to the limit, and the results were similar. If you increase the
scale to 10 1 the air is not enough, so we use oxygen-enriched gas
and in that case we are already working at the limiting conditions.

DIRECT CAPTURE OF MONOCLONAL ANTIBODIES WITH A NEW HIGH
CAPACITY STREAMLINE SP XL ION EXCHANGER

C. PRIESNER, N. AMESKAMP, D. LÜTKEMEYER, J. LEHMANN
Cell Culture Technology, P. O. Box 100131, University of Bielefeld
Germany

1. Abstract

A new sulphopropyl-based strong cation exchange matrix STREAMLINE®SP XL

(Pharmacia Biotech) designed for fluidized bed adsorption was tested in both bench and

pilot scale for direct capture of mouse from undiluted and unclarified feedstock

containing 500 mg/L human serum albumin (HSA) and 30-50 mg/L MAb. In scouting

experiments in packed bed mode with clarified feedstock the optimal pH for binding of
the target protein (pI = 6.3) was found to be 4.6 where co-chromatography of HSA (pI =

4.8) predictably occurred. Frontal adsorption analysis under the same conditions gave
enhanced dynamic capacities of more than 6 mg IgG/ml STREAMLINE®SP XL matrix

compared to those attainable with conventional STREAMLINE®SP matrix (1.5 mg

IgG/ml). These results could not be reproduced in expanded bed mode with unclarified

feedstock (0.5 mg IgG/ml matrix) where massive biomass adsorption to the matrix

could be observed. By adding glucose in different concentrations to the feedstock and

buffers the dynamic capacity increased to 2 mg IgG/mL matrix (200 mM glucose) and

4.5 mg IgG/ml matrix (400 mM glucose), respectively (Table 1). The co-

chromatography of HSA varied according to the glucose content. In pilot scale no

break-through of the antibody could be achieved due to the limiting dynamic capacity.

2. Materials and Methods

2.1. Cultivation

Murine hybridoma cells were cultivated in 5-, 20-, and 100-L-bioreactors in split batch

mode using a culture media with 500 mg/L HSA, 10 mg/L bovine insulin and 8 mg/L
human transferrin. Except those used in the scouting experiments all feedstocks
contained 0.002 % Pluronic F-68® and 0.003 % Antifoam C®. Until the end of each

batch the number of viable cells reached up to with a viability higher than

90%.

429

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 429-431.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

430

2.2. Chromatography

STREAMLINE®SP and SP XL are strong cation exchangers based on highly cross-
linked 6% agarose including an inert quartz core (density 1.2 g/ml, particle size 100 -
300 m). The sulphonate groups of SP XL are coupled to polydextran spacers to obtain
a higher ligand density. Experiments were carried out in packed bed mode with CIO/10
columns / 4 ml adsorbent / FPLC® system, expanded bed mode with a SL®25 column /
75 ml adsorbent / GradiFrac® system (bench scale), and a SL®200 column / 5000 ml
adsorbent / pilot system (pilot scale). All chromatographic equipment: Pharmacia
Biotech.
Buffer A: 20 mM Acetate, 150 mM NaCl, pH 4.6, 0/200/400 mM Glucose; Buffer B:
20 mM Acetate, 1 M NaCl, pH 4.6, 0/200/400 mM Glucose. Equilibration, loading
and washing: 300 cm/h, upward flow direction; Elution: 100 cm/h, upward flow
direction.

2.3 Analytics

Total antibody concentration was measured by ELISA. Particle number and distribution
were measured with a Casyl-system (Schärfe) and interpreted according to the
definition of Glauner [1].

3. Results

431

4. Summary and conclusion
The new high capacity cation exchange matrix STREAMLINE SP XL has proved its
enhanced capacity compared to STREAMLINE SP. The values for dynamic capacities

for pure substances could not be reached due to the co-chromatography of HSA, other
media components, high salt concentration in the feedstock and the biomass adsorption
to the matrix. A scale up was successfully performed: 100 L of unclarified hybridoma
broth were concentrated with a binding rate of 96 % in less than three hours. Future
investigations have to focus on the observed effects of additives as Glucose and
surfactants as well as on the cell binding which seems to be the major problem of the

expanded bed ion exchange chromatography under the chosen conditions.
5. Acknowledgement
We would like to thank Pharmacia Biotech - especially Dr. P. Girard and Dr. L. Larsson
- for kindly supplying the pilot equipment and 5 L of SP XL matrix.

6. References

Glauner, B. (1991) Vitalitätskontrolle in der Zellkulturtechnik; BioTec 3(5)

THE EFFECT OF DIFFERENT PURIFICATION SCHEMES ON THE
ACTIVITY OF A MONOCLONAL ANTIBODY

D.J. BLACK1, J.P. BARFORD1, C. HARBOUR2, and A. FLETCHER3
Departments of Chemical Engineering1 and Infectious Diseases1,
University of Sydney, NSW 2006 and NSW Blood Transfusion Service3,
Sydney, NSW 2000, Australia

Abstract
Five different chromatographic purification schemes, encompassing affinity, hydrophobic
interaction, cation exchange and gel filtration techniques, were developed and used to
purify monoclonal antibody from a common supernatant pool produced by human
lymphoblastoid cells grown in batch culture. Comparison of purified samples revealed that
the purification scheme could affect both the antigen binding and functional activities of
the purified product.

Introduction
Advances in genetic engineering and cell culture technology have made it possible to
generate an increasing number of biopharmaceuticals in non-traditional ways. Factor VIII
for example, a coagulation factor, can now be derived from mammalian cell cultures rather
than the traditional source of pooled human plasma. Many of these biopharmaceuticals are
glycoproteins and it is recognised that the post-translational modifications involved in the
production of complex biological molecules can differ from species to species and thus
result in different product profiles when produced in vitro compared to in vivo. Thus the
product is often a complex mixture of different glycoforms which may have different
activities. It is possible that the use of different purification systems could preferentially
select certain glycoforms resulting in final products with different glycoform compositions
and hence activities. In order to address this issue a human monoclonal antibody with both
antigen binding (Fab-dependent) and functional (Fc-dependent) activities was selected as
the model protein for study. The effects of five different purification schemes on both
these activities have been studied and a preliminary report is presented here.

Materials and methods
Human monoclonals (IgG) with therapeutic potential (anti-D) were produced from
lymphoblastoid cells (generated by Dr A. Fletcher at the NSW BTS) grown in IMDM
medium supplemented with 10% fetal calf serum and 2 mM glutamine. A pooled, single
batch which was then aliquoted and stored at -20°C until purified using one of five different

purification schemes shown in the tables, i.e., protein G; protein A; anti-human IgG-
agarose; two stage hydrophobic interaction chromatography (HiC) and combined HiC, ion
exchange (Ion-X) and gel filtration (GF) chromatography. Antigen (Fab-dependent)

433

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 433-435.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

434