Cytotechnology 9:51-57, 1992. 9 1992 KluwerAcademicPublishers. Printedin the Netherlands.
The m e m b r a n e dialysis bioreactor with integrated radial-flow fixed bed - a n e w approach for continuous cultivation of animal ceils Amain B o h m a n n I, Ralf P6rtner 1, J~Srg Schmieding 2, Volker Kasche 2, Herbert M~irkl 1
ITechnische Universitgit Hamburg-Harburg, BioprozeJ3- und Biovelfahrenstechnik; 2Biotransformation und Biosensorik, DenickestraJ3e 15, 2100 Hamburg 90, Germany Received 1 April 1992; accepted in revised form 24 August 1992
Key words: antibody consistency, c h r o m o s o m e number, continuous cultivation, fixed bed reactor, hybridoma, m e m b r a n e dialysis bioreactor
Abstract A h y b r i d o m a cell was cultivated continuously in a m e m b r a n e dialysis bioreactor with an integrated radial-flow fixed bed consisting of porous Siran | carriers over a period of 6 weeks. Antibodies accumulated to an average of 100 m g 1-l, approx. 10 times more than in fixed bed cultures without dialysis m e m b r a n e . S e r u m costs could be reduced about 85% due to an appropriate feeding strategy. Siran | carriers with 3 - 5 m m diameter showed an advantage c o m p a r e d to those with 1 - 2 m m diameter. For the 3 - 5 m m cartier the specific glucose uptake rate and the M A b production rate were constant, if the velocity was between 0.09 m m s -1 and 0.75 m m s -1. At higher velocities cells are washed out of the bed. Furthermore antibody consistency and cell stability were verified in long-term cultivations over a period of 96 days. F r o m an estimation o f the antibody concentration reachable with the reactor concept under optimal conditions a concentration 45 times higher c o m p a r e d to axial-flow fixed bed reactors and 1 1 times higher c o m p a r e d to stirred tank reactors can be expected.
List of symbols: A c(Glc) c(MAb) c(O2) D U q*(Glc) q*(MAb) V
cross sectional area o f the axial-flow fixed bed glucose concentration monoclonal antibody concentration oxygen concentration dilution rate linear velocity in the axial-flow fixed bed specific glucose uptake rate specific M A b production rate volumetric flow rate in the axial-flow fixed bed
m2 m m o l 1-I m g 1-I % air saturattion d -l m m s -l m m o l I(FB) -1 h -l m g I(FB) -1 h -1 m 3 h -1
52 Introduction
Materials and methods
Among the reactor systems proposed for the cultivation of animal cells fixed bed reactors (Racher, 1990) and reactors with integrated dialysis membranes (Linardos, 1992) are of growing interest. Our membrane dialysis reactor with integrated radial-flow fixed bed (Kurosawa, 1991; P6rtner, 1991) combines these engineering principles: immobilization of cells in a radial-flow fixed bed, which offers a high potential for scaleup, supply of nutrients and removal of toxic metabolites via dialysis membrane and in-process enrichment of products (e.g., monoclonal antibodies). The investigations focussed on important engineering parameters regarding this reactor. Only little is known about the influence of varying flow velocities through a fixed bed on the cell activity. Hybridoma cells, which were used here, are retained inside of a carrier matrix mainly due to mechanisms as in deep bed filtration. Therefore a relationship between the flow velocity and the number of cells retained in the matrix and a wash-out of cells at high velocities can be expected. This information is important, because first in a radial geometry a velocity gradient along the radius exists and second the wash-out velocity limits the length of the bed due to possible oxygen limitation at the outlet. A suitable Siran | carrier for immobilization should be selected. A feeding strategy should be tested, where medium supplemented with serum is added directly to the ceils, while low molecular weight substrates are supplied via the dialysis membrane to save serum costs. The reactor was developed especially for longterm cultivation. For this reason it should be evaluated whether the hybridoma cell line used here and the produced antibody are stable during long-term immobilization. Basic studies regarding flow velocity, antibody consistency and cell stability were performed in a small scale axial-flow fixed bed. The serum feeding strategy was tested in the membrane dialysis reactor with integrated radial-flow fixed bed.
Cell line and culture medium A mouse-mouse hybridoma cell line (Niebuhr-R., 1990) producing monoclonal antibodies (MAb) against penicillin-G-amidase was cultivated in a 1:1 mixture of Iscove's MEM and Ham's F12 supplemented with 3% fetal calf serum (FCS), 2 mmol 1-t L-glutamine and 2 g 1-1 NaHCO3.
Analyses Glucose, lactate, ammonia and L-glutamine were determined enzymatically, monoclonal antibody (MAb) by a mouse-IgG-ELISA (all kits from Boehringer, Germany).
Antibody consistency and chromosome number Antibody consistency was determined by isoelectric focussing (IEF). The culture broth was desalted, concentrated by ultrafiltration and used for IEF. The total protein was transferred electrophoretically on PVDF-membrane. The monoclonal antibody was stained with anti-mouse horseradish peroxidase. The chromosome number (2n) as a marker for stability of the cultured cells was defined by counting 29 to 47 metaphases spreads on slides using solid Giemsa stain.
> G a s Out < G a s In
Inoc. Vessel
Fixed Bed
Reservoir Vessel
-
used
fresh
Medium
Fig. 1. Experimentalset-up of the axial-flowfixedbed reactor.
53
Carrier material Fractions of porous Siran | glass carriers (Schott, Germany) with mean diameters o f I - 2 m m and 3 - 5 m m were used for immobilization of the hybridoma cells, giving an external void fraction of 0.3 in the fixed bed. The pore size is less than 120 I.tm for the 1 - 2 m m fraction and 6 0 - 3 0 0 g m for the 3 - 5 m m fraction, the estimated surface areas are 87 m 2 1-1 and 74 m 2 1-I, respectively. The pore v o l u m e o f the carrier is approx. 55--60% in both cases (Looby, 1989).
40 mm) containing Siran| carriers ( I - 2 mm) is arranged in the inner chamber. Fresh medium and oxygen are supplied from the outer chamber. Toxic metabolites such as lactic acid and ammonia are dialysed through the membrane, while antibodies are retained and accumulated in the inner chamber. For continuous cultivation fresh medium was perfused through the outer chamber.The inner chamber was operated in a semi-batch mode.
A.xial-flow fixed bed for basic studies An axial-flow fixed bed (volume 50 ml, diameter 40 mm, length 40 mm, Fig. 1), connected with a 750 ml reservoir vessel for aeration and medium exchange, was used for basic studies focussing on flow velocity, antibody consistency and cell stability. The medium in the reservoir vessel could be changed in a semi-batch mode or continuously. Dissolved oxygen (DO) was controlled in the reservoir vessel and at the outlet of the fixed bed. The fixed bed matrix was build up with 3 - 5 m m Siran | carriers. The length o f the bed had been chosen in order to prevent oxygen limitation at the outlet (P6rtner, 1991), as could be confirmed during the experiment. For inoculation the inoculation vessel was filled with a cell suspension o f exponentially growing cells harvested from standard culture flasks. The cell suspension was p u m p e d through the fixed bed from the bottom and from the top of the bed alternately to get a uniform cell distribution throughout the whole carrier matrix.
r membr
fixed d
Membrane dialysis bioreactor with integrated radial-flow fixed bed The dialysis bioreactor (Kurosawa, 1991) consists o f two chambers (outer chamber 6 1, inner chamber 1 1), which are separated by a dialysis membrane (Cuprophan| Enka, Germany) with cut off 10,000 (Fig. 2). The radial-flow fixed bed (volume 600 ml, inner radius 5 ram, outer radius
ouler / chomber
~
~
"
~inner ch.omber
Fig.2. Membranedialysis bioreactorwith radial-flowfixed bed.
54 Inoculation was done by adding a cell suspension to the inner chamber. The cell suspension was p u m p e d through the bed by the centrifugal p u m p with a low rotation speed until the m e d i u m b e c a m e clear.
~"
121
I ......... t ....................
io-,~ I
E_ "
Basic studies in the axial-flow fixed bed focussed on the influence of varying linear velocities U, defined as the volumetric flow rate "Q through the bed divided by the cross sectional area A of the bed, on the volumetric cell activity. The question was, whether an influence of the velocity on the volumetric cell activity due to shear stress has to be considered. For this non-adherent, but slightly sticky cell line it can be assumed that with increasing velocity a growing number of cells will be washed out the bed resulting in a decreasing volumetric cell activity. Furthermore it should be evaluated, if there is a possible change of the antibody productivity and the antibody consistency or the c h r o m o s o m e n u m b e r of the cells during long-term immobilization. The axial-flow fixed bed (FB) reactor was run for 96 days (Fig. 3) by varying linear flow velocities U. The bed was inoculated with I x 106 cells mI(FB) -l. During the cultivation the oxygen level in the reservoir vessel was held between 90 and 100% of air saturation. Up to day 45 tile reactor was run in a semi-batch m o d e in the reservoir vessel with an exchange of 80% m e d i u m every day. Between day 45 and day 96 the reservoir vessel was perfused with a dilution rate of 1 d -I. As an indicator for cell activity the specific glucose uptake rate q*(Glc) is shown in Fig. 3. The glutamine uptake was proportional to the glucose uptake (0.14 tool glutamine per tool glucose, data not shown). During the first 13 days the m e d i u m velocity through the bed was gradually increased to find the wash-out velocity. On day 14 at U = 1.1 m m s -I a drastic increase of suspended cells in the
$~'anr
" ~
3-5 rnrn
0
Results and discussion
Basic studies in the crt'ial-flow fixed bed
t
1.2
'= 0.8
_
_
_
U
0
O~
0
1
,
"
.
2O
0
0
,
20
-
,
,
40
60
,
80
100
time [dJ
Fig. 3.
C u l t i v a t i o n in the a x i a l - f l o w f i x e d b e d at d i f f e r e n t l i n e a r
v e l o c i t i e s U. S p e c i f i c g l u c o s e u p t a k e rate q * ( G l c ) , l i n e a r v e l o c ity U , o x y g e n c o n c e n t r a t i o n at the o u t l e t a n d M A b c o n c e n t r a t i o n
c(MAb) vs. time.
reservoir vessel was observed and the glucose uptake rate dropped, indicating a wash-out o f cells. At U = 0.6 m m s -l the cultivation could be maintained. After 27 days the velocity was increased in smaller steps and at U = 0.96 m m s -I wash-out occurred (day 32). From day 46 to day 96 the linear velocity was varied between 0.75 and 0.09 m m s -l. Even at this low velocity the oxygen concentration at the outlet was still higher than 38% of air saturation. Each velocity was held constant for at least 3 days. A small n u m b e r of cells was permanently washed out of the bed and suspended in the
55 150"
12
l Siran carrier 3-5 mrnJ
"7
t
X X
100-
x total
z viable I
X X
"7 i,-
X
(,9
/
X
"7
"7
,,'P 4-
g
E El. (/} ::3
X
0'.2
X
0'.4
0'.6
U
0'.8
[ mms" ]
X
8
•
~" u_ v
O'J
E E
E X
D
The m e m b r a n e dialysis bioreactor with integrated radial-flow fixed bed - a n e w approach for continuous cultivation of animal ceils Amain B o h m a n n I, Ralf P6rtner 1, J~Srg Schmieding 2, Volker Kasche 2, Herbert M~irkl 1
ITechnische Universitgit Hamburg-Harburg, BioprozeJ3- und Biovelfahrenstechnik; 2Biotransformation und Biosensorik, DenickestraJ3e 15, 2100 Hamburg 90, Germany Received 1 April 1992; accepted in revised form 24 August 1992
Key words: antibody consistency, c h r o m o s o m e number, continuous cultivation, fixed bed reactor, hybridoma, m e m b r a n e dialysis bioreactor
Abstract A h y b r i d o m a cell was cultivated continuously in a m e m b r a n e dialysis bioreactor with an integrated radial-flow fixed bed consisting of porous Siran | carriers over a period of 6 weeks. Antibodies accumulated to an average of 100 m g 1-l, approx. 10 times more than in fixed bed cultures without dialysis m e m b r a n e . S e r u m costs could be reduced about 85% due to an appropriate feeding strategy. Siran | carriers with 3 - 5 m m diameter showed an advantage c o m p a r e d to those with 1 - 2 m m diameter. For the 3 - 5 m m cartier the specific glucose uptake rate and the M A b production rate were constant, if the velocity was between 0.09 m m s -1 and 0.75 m m s -1. At higher velocities cells are washed out of the bed. Furthermore antibody consistency and cell stability were verified in long-term cultivations over a period of 96 days. F r o m an estimation o f the antibody concentration reachable with the reactor concept under optimal conditions a concentration 45 times higher c o m p a r e d to axial-flow fixed bed reactors and 1 1 times higher c o m p a r e d to stirred tank reactors can be expected.
List of symbols: A c(Glc) c(MAb) c(O2) D U q*(Glc) q*(MAb) V
cross sectional area o f the axial-flow fixed bed glucose concentration monoclonal antibody concentration oxygen concentration dilution rate linear velocity in the axial-flow fixed bed specific glucose uptake rate specific M A b production rate volumetric flow rate in the axial-flow fixed bed
m2 m m o l 1-I m g 1-I % air saturattion d -l m m s -l m m o l I(FB) -1 h -l m g I(FB) -1 h -1 m 3 h -1
52 Introduction
Materials and methods
Among the reactor systems proposed for the cultivation of animal cells fixed bed reactors (Racher, 1990) and reactors with integrated dialysis membranes (Linardos, 1992) are of growing interest. Our membrane dialysis reactor with integrated radial-flow fixed bed (Kurosawa, 1991; P6rtner, 1991) combines these engineering principles: immobilization of cells in a radial-flow fixed bed, which offers a high potential for scaleup, supply of nutrients and removal of toxic metabolites via dialysis membrane and in-process enrichment of products (e.g., monoclonal antibodies). The investigations focussed on important engineering parameters regarding this reactor. Only little is known about the influence of varying flow velocities through a fixed bed on the cell activity. Hybridoma cells, which were used here, are retained inside of a carrier matrix mainly due to mechanisms as in deep bed filtration. Therefore a relationship between the flow velocity and the number of cells retained in the matrix and a wash-out of cells at high velocities can be expected. This information is important, because first in a radial geometry a velocity gradient along the radius exists and second the wash-out velocity limits the length of the bed due to possible oxygen limitation at the outlet. A suitable Siran | carrier for immobilization should be selected. A feeding strategy should be tested, where medium supplemented with serum is added directly to the ceils, while low molecular weight substrates are supplied via the dialysis membrane to save serum costs. The reactor was developed especially for longterm cultivation. For this reason it should be evaluated whether the hybridoma cell line used here and the produced antibody are stable during long-term immobilization. Basic studies regarding flow velocity, antibody consistency and cell stability were performed in a small scale axial-flow fixed bed. The serum feeding strategy was tested in the membrane dialysis reactor with integrated radial-flow fixed bed.
Cell line and culture medium A mouse-mouse hybridoma cell line (Niebuhr-R., 1990) producing monoclonal antibodies (MAb) against penicillin-G-amidase was cultivated in a 1:1 mixture of Iscove's MEM and Ham's F12 supplemented with 3% fetal calf serum (FCS), 2 mmol 1-t L-glutamine and 2 g 1-1 NaHCO3.
Analyses Glucose, lactate, ammonia and L-glutamine were determined enzymatically, monoclonal antibody (MAb) by a mouse-IgG-ELISA (all kits from Boehringer, Germany).
Antibody consistency and chromosome number Antibody consistency was determined by isoelectric focussing (IEF). The culture broth was desalted, concentrated by ultrafiltration and used for IEF. The total protein was transferred electrophoretically on PVDF-membrane. The monoclonal antibody was stained with anti-mouse horseradish peroxidase. The chromosome number (2n) as a marker for stability of the cultured cells was defined by counting 29 to 47 metaphases spreads on slides using solid Giemsa stain.
> G a s Out < G a s In
Inoc. Vessel
Fixed Bed
Reservoir Vessel
-
used
fresh
Medium
Fig. 1. Experimentalset-up of the axial-flowfixedbed reactor.
53
Carrier material Fractions of porous Siran | glass carriers (Schott, Germany) with mean diameters o f I - 2 m m and 3 - 5 m m were used for immobilization of the hybridoma cells, giving an external void fraction of 0.3 in the fixed bed. The pore size is less than 120 I.tm for the 1 - 2 m m fraction and 6 0 - 3 0 0 g m for the 3 - 5 m m fraction, the estimated surface areas are 87 m 2 1-1 and 74 m 2 1-I, respectively. The pore v o l u m e o f the carrier is approx. 55--60% in both cases (Looby, 1989).
40 mm) containing Siran| carriers ( I - 2 mm) is arranged in the inner chamber. Fresh medium and oxygen are supplied from the outer chamber. Toxic metabolites such as lactic acid and ammonia are dialysed through the membrane, while antibodies are retained and accumulated in the inner chamber. For continuous cultivation fresh medium was perfused through the outer chamber.The inner chamber was operated in a semi-batch mode.
A.xial-flow fixed bed for basic studies An axial-flow fixed bed (volume 50 ml, diameter 40 mm, length 40 mm, Fig. 1), connected with a 750 ml reservoir vessel for aeration and medium exchange, was used for basic studies focussing on flow velocity, antibody consistency and cell stability. The medium in the reservoir vessel could be changed in a semi-batch mode or continuously. Dissolved oxygen (DO) was controlled in the reservoir vessel and at the outlet of the fixed bed. The fixed bed matrix was build up with 3 - 5 m m Siran | carriers. The length o f the bed had been chosen in order to prevent oxygen limitation at the outlet (P6rtner, 1991), as could be confirmed during the experiment. For inoculation the inoculation vessel was filled with a cell suspension o f exponentially growing cells harvested from standard culture flasks. The cell suspension was p u m p e d through the fixed bed from the bottom and from the top of the bed alternately to get a uniform cell distribution throughout the whole carrier matrix.
r membr
fixed d
Membrane dialysis bioreactor with integrated radial-flow fixed bed The dialysis bioreactor (Kurosawa, 1991) consists o f two chambers (outer chamber 6 1, inner chamber 1 1), which are separated by a dialysis membrane (Cuprophan| Enka, Germany) with cut off 10,000 (Fig. 2). The radial-flow fixed bed (volume 600 ml, inner radius 5 ram, outer radius
ouler / chomber
~
~
"
~inner ch.omber
Fig.2. Membranedialysis bioreactorwith radial-flowfixed bed.
54 Inoculation was done by adding a cell suspension to the inner chamber. The cell suspension was p u m p e d through the bed by the centrifugal p u m p with a low rotation speed until the m e d i u m b e c a m e clear.
~"
121
I ......... t ....................
io-,~ I
E_ "
Basic studies in the axial-flow fixed bed focussed on the influence of varying linear velocities U, defined as the volumetric flow rate "Q through the bed divided by the cross sectional area A of the bed, on the volumetric cell activity. The question was, whether an influence of the velocity on the volumetric cell activity due to shear stress has to be considered. For this non-adherent, but slightly sticky cell line it can be assumed that with increasing velocity a growing number of cells will be washed out the bed resulting in a decreasing volumetric cell activity. Furthermore it should be evaluated, if there is a possible change of the antibody productivity and the antibody consistency or the c h r o m o s o m e n u m b e r of the cells during long-term immobilization. The axial-flow fixed bed (FB) reactor was run for 96 days (Fig. 3) by varying linear flow velocities U. The bed was inoculated with I x 106 cells mI(FB) -l. During the cultivation the oxygen level in the reservoir vessel was held between 90 and 100% of air saturation. Up to day 45 tile reactor was run in a semi-batch m o d e in the reservoir vessel with an exchange of 80% m e d i u m every day. Between day 45 and day 96 the reservoir vessel was perfused with a dilution rate of 1 d -I. As an indicator for cell activity the specific glucose uptake rate q*(Glc) is shown in Fig. 3. The glutamine uptake was proportional to the glucose uptake (0.14 tool glutamine per tool glucose, data not shown). During the first 13 days the m e d i u m velocity through the bed was gradually increased to find the wash-out velocity. On day 14 at U = 1.1 m m s -I a drastic increase of suspended cells in the
$~'anr
" ~
3-5 rnrn
0
Results and discussion
Basic studies in the crt'ial-flow fixed bed
t
1.2
'= 0.8
_
_
_
U
0
O~
0
1
,
"
.
2O
0
0
,
20
-
,
,
40
60
,
80
100
time [dJ
Fig. 3.
C u l t i v a t i o n in the a x i a l - f l o w f i x e d b e d at d i f f e r e n t l i n e a r
v e l o c i t i e s U. S p e c i f i c g l u c o s e u p t a k e rate q * ( G l c ) , l i n e a r v e l o c ity U , o x y g e n c o n c e n t r a t i o n at the o u t l e t a n d M A b c o n c e n t r a t i o n
c(MAb) vs. time.
reservoir vessel was observed and the glucose uptake rate dropped, indicating a wash-out o f cells. At U = 0.6 m m s -l the cultivation could be maintained. After 27 days the velocity was increased in smaller steps and at U = 0.96 m m s -I wash-out occurred (day 32). From day 46 to day 96 the linear velocity was varied between 0.75 and 0.09 m m s -l. Even at this low velocity the oxygen concentration at the outlet was still higher than 38% of air saturation. Each velocity was held constant for at least 3 days. A small n u m b e r of cells was permanently washed out of the bed and suspended in the
55 150"
12
l Siran carrier 3-5 mrnJ
"7
t
X X
100-
x total
z viable I
X X
"7 i,-
X
(,9
/
X
"7
"7
,,'P 4-
g
E El. (/} ::3
X
0'.2
X
0'.4
0'.6
U
0'.8
[ mms" ]
X
8
•
~" u_ v
O'J
E E
E X
D
