Appl Microbiol Biotechnol (1991) 35:165-175

Applied ., Microbiology Biotechnology

0175759891001095

© Springer-Verlag 1991

Some aspects of hybridoma cell cultivation H. Graf I and K. Schiigerl 2 1 BASF, Ludwigshafen, Federal Republic of Germany 2 Institut for Technische Chemie, Universit~it Hannover, Callinstrasse 3, W-3000 Hannover, Federal Republic of Germany Received 30 November 1990/Accepted 17 December 1990

Summary. T w o h y b r i d o m a cell lines were cultivated in an indirectly aerated 10-1 reactor in batch, f e d - b a t c h a n d c o n t i n u o u s (perfusion) operations a n d in spinner flasks. The m e d i u m in the reactor was s a m p l e d either by an aseptic cross-flow filtration m o d u l e integrated into a l o o p or b y an in-situ t u b u l a r filter. The glucose c o n c e n t r a t i o n was m o n i t o r e d by an on-line flow injection analyser a n d the a m m o n i a c o n c e n t r a t i o n b y an ion-selective electrode. Since the m e m b r a n e transmission o f the h i g h - m o l e c u l a r c o m p o n e n t s decreased during cultivation, the product, a m o n o c l o n a l antibody, was e n r i c h e d in the reactor. D u r i n g cultivation, the c o n c e n t r a t i o n s o f cells, viable cells, glucose, lactase, acetate, citrate, a m m o n i a , urea, a m i n o acids, proteins, a n d m o n o c l o n a l antibodies were d e t e r m i n e d off-line. The specific g r o w t h rate, specific p r o d u c t i o n , a n d cons u m p t i o n rates o f the m e d i u m c o m p o n e n t s were inf l u e n c e d c o n s i d e r a b l y b y the m e d i u m c o m p o s i t i o n , especially b y the type a n d a m o u n t o f serum used.

Introduction Culture m e d i u m c o m p o s i t i o n has considerable influence o n m a m m a l i a n cell g r o w t h a n d p r o d u c t f o r m a tion (e.g., G l a c k e n et al. 1986; Miller et al. 1989a, b; W a g n e r et al. 1987). On-line m o n i t o r i n g a n d control o f m e d i u m c o m p o n e n t s in microbial cultivation m e d i a are n o w c o m m o n p l a c e on a l a b o r a t o r y scale (Schiigerl 1988). H o w e v e r , except for temperature, pO2, p H , a n d stirrer speed, no other process variables are controlled by in-situ or on-line m e a s u r e m e n t s (e.g., Wallberg et al. 1987). There are only a few publications on the application o f on-line m o n i t o r i n g for m a m m a l i a n cell cultivation (Schiigerl et al. 1990). The present p a p e r reports on the d e v e l o p m e n t a n d application o f substrate c o n c e n tration control during m a m m a l i a n cell cultivation.

Offprint requests to: K. Schiigerl

Materials and methods Cell lines. Two mouse-mouse cell lines (F34 and 3C2) were prepared by P. Nabet (personal communication) by the fusion of myeloma cells of cell line P3X63-Ag8 with activated B-lymphocytes of a mouse. The mouse-mouse 3C2 cells produce monoclonal antibody (MAB) immunoglobulin G-1 (IgG-1) against the hormone Human-Choriongonadotropin. The cell line F34 does not produce antibodies. The hybridoma cell stock cultures were stored at 37 ° C and 5% CO2 in air in Roux flasks (Falcon, Cockeyville, USA) in incubators (Heraeus Type B 5060 EK CO2; Hanau) and were diluted twice a week with fresh medium to 105 cells ml h -1. These cells were used as seeds for the precultures in Bellco (Vineland, N J, USA) spinner flasks. Culture media. The basic medium for F34 and 3C2 was RPMI 1640 powder (Gibco, Grand Island, USA), supplemented with 0.328 g 1-1 glutamine (Serva, Heidelberg, FRG), 0.132 g 1-1 sodium pyruvate (Riedel de Haen, Seelze), 10 lxg 1-1 2-mercaptoethanol (Serva), and 2.0 g 1-1 NaHCO3 (Riedel de Haen). For cultivation in the 10-1 reactor 0.060 g 1-I gentamicin (Serva) in distilled water (ASTM-Type 1, R = 18 M~ cm -1) was also added. Before inoculating the medium, it was sterile-filtered, and various amounts of either foetal calf serum (FCS; Gibco) or horse serum (HS; Gibco) was added to the medium. The pH value was controlled by addition of gaseous CO2 or 0.1 N NaOH. Bioreactor. A 10-1 bioreactor (Biostat E, Braun, Melsungen, FRG) with a low stirrer speed (< 200 rpm), 15.2 m silicone tubing (3 mm diameter and 0.4 mm wall thickness) wound around a cylindrical basket of 16 cm diameter for indirect aeration and 10 m hydrophilized microporous polytetrafluorethylene (PTFE) tubing (Gore, Putzbrunn, FRG) (2 mm diameter 0.4 mm wall thickness) for medium exchange with perfusion, was used for cultivation (Fig. 1). The oxygen transfer rate and pH were controlled by the gas composition (N2:O2:CO2) in the silicone tubing. On-line analysis. Two aseptic sampling systems were used: (a) a cross-flow flat filtration module (Millipore, Freehold, N J, USA) (Fig. 2) with a pump (Watson-Marlow 101 UR, Eschborn, FRG) in the outer loop, and (b) a tubular filter developed in the Technical Chemistry Institute (TCI) at the University of Hannover and sold by ABC (Puchheim, FRG) (Fig. 3). In the Millipore module, a hydrophilic flat membrane (GVWP 04700, Millipore) was used. In the tubular filter, polypropylene microfiltration tubing (Enka, Wuppertal, FRG) was used. The sampling modules are characterized in Table 1.

166 off-line ~

sampling

F1

on-line,0m~ inoculum ~ V1 F2

-- V 2

-

V

....

V4

-[~~~, ~

t

l

exhaustgas

t !

,Ill X ~

"11

I~ J! I!

Pl

.......

"~"

P2

ll dischargestorage

~._~

~

....... 1~

~

~o~ ~ o~ ~ra~e

....... ~...................

stirrer basket for PTFE and silicon tubing,

mediumstorage

V

~

steam

Fig. 1. Experimental set-up: P, pump; V, valve; F, membrane filter; PTFE, polytetrafluorethylene

Connection tube

Connectionpiece active surface area 34.5 cm 2 dead volume 7.0 rnl response time 9,0 rain mean pore diameter 0,2~m

retentateside

filtrateside ,

II

Fig. 2. Cross-flow flat membrane filtration module for on-line aseptic sampling The on-line flow injection analyser (FIA) system for glucose analysis consisted of a 16-channel peristaltic pump (Skalar, Erkelenz, FRG) tygon-tubing, motor valve (Latek-TMV, Eppenheim, FRG), injector valve (Lee Hydraulic Miniature Components, Frankfurt, FRG) and a YSI-glucose analyser (Model 23A, Yellow Spring Instruments, Ohio, USA) modified for on-line operation (TCI, Hannover). The operation of the FIA system was controlled by a microprocessor (Motorola, Type 68 000) and a suitable software package (FERAS; Wieneke 1989).

Extensionpiece

~amplingmodule ~ith nicrofiltration nembrane

Samplingmodule without microfiltration membrane

Fig. 3. Tubular filtration module for in-situ sampling

167

L::'"2:+

Table 1. Properties of the sampling modules Parameters

Flat Millipore module

Tubular Enka module

Free filtration area (cm 2) Dead volume on the permeate side (ml) Response time (min) Membrane material

7.2

30

1.0 1.0 Polyvinylidene fluoride

8 9 Polypropylene

Mean membrane thickness (mm) Mean membrane pore diameter (txm)

reagent - -

0.125

co 2 NaOH N2

O2

1.5

0.22

,,

D/D out [ ]

[] [] [] []

0.2

~

i i i i i,i i,l,t,i

ammonia electrode

25oc

D/Do o t - [ ]

E

reagent 1 ~ H20 air reagent

2

--

/'

90oc

27oc

:0nm

--

i---

multichannel peristalticpump Fig. 4. Continuous flow analysers for ammonia and reducing sugar (glucose monitors)

Reducing sugar was determined in a continuous air-segmented analyser (Skalar) by means of p-hydroxybenzoic acid hydrazide (p-HBAH) (Schmidt et al. 1985) and a photometer at 410 nm (Type V5, Julabo Labortechnik, Seelbach, FRG) (Fig. 4). The determination of the ammonia concentration was carried out by an ion-selective electrode (Type 9512, Orion Research, Cambridge, Mass., USA) and the pO2 by an 02 electrode (Bauermeister 1978). The transfer, control, and evaluation of on-line analysis data were performed by a computer system (VME 68 000/RTOS operating system, and PDP ll/23+/RSX-11M + operating system, Fig. 5) as well as by FERAS and CASFA (Dors 1989) software packages.

Off-line analysis. The cell concentration was determined by a Thoma chamber, the vitality with trypan blue, the concentration of glucose by off-line FlA. The following off-line analyses were carried out: the Giuco-DH method (Mercotest 14335, Merck, Darmstadt, FRG), lactate (Boehringer, Mannheim, FRG; BM 149993), urea (urease), citrate (BM 139076), acetate (BM 14826), ammonia (BM 542 946). The analysis of protein was performed by the Bicinchonin acid-method of Pierce (Smith et al. 1985), the amino acids by the HPLC-o-phthaldialdehyde method (Kretzmer 1986) and the MAB with sandwich ELISA (Nabet, personal communication).

samplingvalve magneticvalve for rinsingsolution magneticvalve for glucosestandard feed pump

~00 "

.i/

D/DIn - - ~

terminal

sample

feed pump harvest pump

[]

samplingvalve liquidlevel measure/control dissolved oxygenmeasure/control pH measureand control

D/D in

~

temperature measure/control stirrer speed glucose analysis ammoniaanalysis

Fig. 5. Schematic outline o~ the dat~ monitoring and process con-

t~o] s~stcm: ~ / ~ , ~J~it~]/~Jta]

Results

Determination of specific consumption rates of medium components by cell line F34 NP The cells were cultivated in RPMI 1640 medium, supplemented with 2.5 g 1-1 glucose, NaHCO3, Na-pyruvate, 2-mercaptoethanol, glutamine, gentamicin and 4% HS at 37°C and pH 7.2 in batch operation with an initial cell concentration of 6.104 cells m l - 1. After an 18 h lag phase, 100 h exponential growth (specific growth rate,/z = 0.0278 h - 1), and a 62 h stationary phase, a viable cell concentration of 1.02.10 -6 cells m1-1 was attained. In Fig. 6, the viable cell concentration and oxygen uptake rate (OUR) are shown as a function of the cultivation time. They attained a maximum at 125 h and/or 150 h and had similar courses. The specific substrate consumption rates during exponential growth are shown in Table 2. The essential amino acids (arginine, leucine, valine, threonine, isoleucine, methionine and lysine) were consumed at fairly high rates. It was not possible to determine histidine because its small peak appeared together with the high peak of glutamine in the HPLC spectrum. During fed-batch cultivation with 3 g 1-1 glucose, 4% HS and an initial cell concentration of 5.104 viable cells m1-1 after a 20 h lag, 90-100 h exponential growth (/J = 0.026-0.0285 h - l ) and the stationary phase, 1.3-106 viable cells m1-1 were attained. In Fig. 7, the

168 o;a

~~_ . . . . . . . . ~. .

~

o

f

-

\

/

;/.

,,,'/'

6

,,F"

~

/

"7

%

///f/

~' -~x,~ x ~ ~----~Nx

©

'~ ~

~m

o

~

~o

too t.~0 cultivation time t h;

~o0

~

2s0

s00

line

Table 2. Specific substrate consumption rates (qs) in nM h)-1 of F34 cells during the exponential growth phase

(10 6 cells

Substrate

qs

Substrate

qs

Oxygen Glucose Lactate Aspartate Glutamate Asparagine Serine Glutamine Glycine Threonine

125.30 70.90 489.80 1.48 - 0.88 1.83 1.62 59.60 - 1.79 2.11

Arginine Alanine Tyrosine Methionine Valine Tryptophan Phenylalanine Isoleucine Leucine Lysine

11.20 - 12.50 0.66 1.40 2.28 0.28 0.52 1.76 3.t3 1.01

The + symbol indicates consumption and the - symbol indicates production

•

•

•

•

•

~B~.~

//m/ .//~j~///~/

"'

~

~

,g

#

4

4~

~__ ~.~n

~,~ oma8 ~ ~ ~

~. ~../~

~0

5O

I

~:0

50

o

I 100

I 7.50

2010

I 250

300 °

cultivation time [h ~

~i8, ~, Viable cell concentration and oxygen consumption rate as a function of the cultivation time. Batch cultivation of F34NP cell

•

o

I00

ISO

~ ~0M

~ ~

¢~lfiwfio~ fim~ [h]

Fig. 7. Variation in the logarithm of the viable cell concentration and the glucose concentration measured enzymatically off-line and with the p-hydroxybenzoic acid hydrazide (p-HBAH) method on-line as functions of the cultivation time. The F34 cell line was used with RPMI 1640 medium

Fig. 8. Logarithm of viable cell concentration and oxygen consumption rate as a function of the cultivation time. Batch cultivation of F34 cell line

logarithm of the viable cell concentration, and the glucose concentration measured on-line by p - H B A H and off-line by enzyme FIA are shown as a function of the cultivation time. After about 125 h, the stationary phase was attained. Glucose consumption was fairly high during the stationary phase. After 200 h, the glucose was consumed. In Fig. 8, again the logarithm of the viable cell concentration and OUR, as well as the specific oxygen utilization rate qo2 are shown. During the exponential phase, the qo2 was nearly constant [4.5 ~tg 0 2 (10 6 cells h ) - 1]. A comparison of the results of batch and fed-batch cultivations shows no dramatic differences. Only the lag phase was longer in the fed-batch culture. However, this deviation could be caused by differences in the inoculum quality. After 6 min direct aeration of the cell-containing culture medium, the viable cell concentration dropped to 50% of its original value. No dead cells, but only cell fragments were present. With increasing column height/diameter ratio, the viable cell concentration increased (J~mmrich 1988).

Cultivation o f mouse-mouse hybridomas 3C2 Batch cultivations were performed with RPMI 1640 medium containing 10 and 5% FCS as well as 4% HS. The initial cell concentrations were 5.104 cells m1-1 and the dissolved oxygen saturation was 50%. During the exponential phase, the viable cell fraction was 100%. With 10% FCS after 95 h, the cell viability began to diminish, by 107 h it was only 66% and by 150 h had decreased to below 10%. The rapid reduction in the viable cell fraction after depletion of the glucose is typical for culture media of low amino acid content. The glutamine content was already exhausted at 90 h, i.e., 10 h before the glucose was spent.

169

& ,.,

i

/~

o

o

~~ I"1/ ~ !

~

'7

./

)

./

o

~

'~~.~.~/"'~"

~'~ _~.--.~.---""'-"" ,

,o'o

,;o

20~

~

o ~

cultivation time

[h]

Fig. 9. Viable and total cell concentrations as well as immunoglobulin G (IgG) concentration as a function of the batch cultivation time of 3C2 cells

[]

~

o

I00

~0

J -

~

~

.

20~°

IS0

cultivation time [h] Fig. I0. Concentrations of glucose and lactate as a function of the

batch cultivation time of 3C2 cells

~

~'o C)

I

i

o

/"

,,:\

/

)

\\~',,

/

; \ ""li,.,~ ' l i , ."~ ~\

SO

100

1

isf

I ~0

~

~

~

~o~

~

cultivation time [h]

cultivation time [h]

Fig. 11. Oxygen consumption rate calculated from the dissolved oxygen concentrations measured at the top and in the bottom (B) of the bioreactor as a function of the batch cultivation time of 3C2 cells

Fig. 12. Glucose consumption rate (GUR) and lactate production rate (LPR) as functions of the batch cultivation time of 3C2 cells

In Fig. 9, the cell and viable cell concentrations as well as the IgG concentration are shown as a function of the cultivation time. After 75 h, the difference between overall and viable cell concentration became apparent, and above 100 h, the deviation increased quickly. The MAB concentration attained its maximum at 125 h. A comparison of Fig. 9 and Fig. 10, in which glucose and lactate concentration courses are shown, indicates the close relationship between the glucose concentration and the reduction in the viable cell fraction. A comparison of Fig. 9 and Fig. 11 again demonstrates the close relationship between the viable cell concentration and OUR. On the other hand, Fig. 12 indicates that most of the glucose was converted into lactate.

Table 3. Influence of the serum type and concentration on the growth and production of mouse-mouse-hybridoma 3C2 in RPMI 1640 medium Parameters

Lag phase (h) /~ (h -1) Max. MAB conc (mg 1-1) Max. viable cell conc (10 6 cells m1-1)

Biostat E

Spinner

10%FCS

5%FCS

4%HS

13 0.04

13 0.033

15 0.035

56

15

19

1.01

0.95

0.58

10%FCS 5 0.039 30 0.8

MAB, monoclonal antibody; FCS, foetal calf serum; HS, horse serum

170

Comparison of batch cultures of 3C2 with FCS and HS

Cyclic batch cultivation of 3C2

Cell line 3C2 has been adapted during several years to 10% FCS. Therefore, it was expected that the medium with 10% FCS would give a better performance than that with HS. In Table 3, batch cultures in a spinner flask with 10% FCS, and in Biostat E with 10%, 5% FCS and 4% HS are compared. Cultures in a spinner flask with 10% FCS had the shortest lag phase and yielded the highest MAB productivity. The culture performance in Biostat E with 10% FCS was better than with 5% FCS and 4% HS, but unexpectedly, 4% HS gave better results than 5% FCS.

The supplemented RPMI 1640 medium was used with 4% HS. The initial cell concentration of 5. | 0 4 cells m1-1 increased with # = 0.036 h-1 to 6. l0 s viable cells ml-1 after 100 h. The medium exchange was performed as soon as the glucose concentration dropped below 0.1 g 1-1 (Fig. 13). The volume was reduced from 10 1 to 5 1 and made up with fresh medium to 10 1. In general, after each cycle, a viable cell concentration of 1.10 -6 and MAB concentration of 20 mg 1-1 were attained (Fig. 14). The lactate concentration variation was a mirror image of the glucose concentration course (Fig. 15). With a sampling flow rate of 0.4 ml min-1 and a volume of the tubular sample filter of 8 ml, the response time of the sampling system was 20 min. The off-line and on-line measured glucose concentrations agreed

&

o

8

~ 8

°

1!

I

100

~0

i~0

•

~

Fig. 13. Variation in glucose concentration during cyclic batch cultivation of the 3C2 cell line. Glucose was determined by three different methods: flow injection analysis (Yellow Springs Instruments, YSI), reducing sugar analysis (on-line) and enzymatic analysis

°

2~0~

200

cultivation time [h]

~'o

~.~

7-'-'-'----\

/"

l..--'~.. . . . ..-- ~,,." ~,~

°IX i

i

:"

"

o/X / 7

o

~.,,.

l/t/t/t/ - W '~

•

E

,

°~ g

~

~

~

/

/

M

i

w

!

4

~ # _.

~ I

~0

-

•

~

ii

¢

tl.

•

..

d

i

./~ ,: W '~

~

_

~ .~ V ~ ~.~

~ " ~

~0

100

1~0

cultivation time [h]

200

2~0

30 0~

Fig. 14. Logarithm of the viable cell concentration, viable cell concentration and IgG concentration as a function of the cultivation time during cyclic batch cultivation of the 3C2 cell line

171 LO~

a~ o

o~1

t~. , I O o

~O ~ • ~ v-~

..

-

o| •

g

o

o

°0

~0

I

l

/

I00

IS0

200

viable cells h) -1. The MAB concentration increased to 60 mg l-1 and attained a maximum level of 90 mg 1-1, an average MAB production rate of 456 ng (10 6 cells h) -1, which was only 10% lower than that in the exponential growth phase. Glucose was partly converted into lactate, and during the oxidation of glutamine, ammonia was formed (Fig. 16). The lactate concentration attained a maximum of 12.1 mM 1-1 a t t - - 134 h, and during the stationary phase it decreased to 7 mM 1-1, probably due to a change in metabolism after the reduction in the glucose concentration. The reduction in the lactate concentration (Fig. 16) and glutamine concentration was accompanied by a rise in the alanine and glycine concentrations (Fig. 17). The amino acids present in the culture medium can be divided into four groups: 1. Amino acids formed: alanine and glycine (Fig. 17). 2. Amino acids consumed at a high rate: leucine, isoleucine, valine, lysine, threonine and methionine, in this order at a decreasing rate (Figs. 18, 19, 20). 3. Amino acids consumed at a low rate: aspartate, phenylalanine and tyrosine (Figs. 18, 20). 4. Amino acids with nearly constant concentrations: serine, aspartate and glutamate. In Table 4 is shown the concentrations and specific consumption rates of the amino acids and metabolites by 3C2 cells during the exponential growth and the stationary phase. This run was repeated with a higher initial cell concentration (1.10 6 viable cells ml-1). At a perfusion rate of 4 1 per day at a steady state, a viable cell concentration of 1.87-10 6 cells m1-1 at a constant ammonia concentration of 30-40 mg 1-1 (Fig. 21) was attained. The specific growth rate ~ = 0.0153 h - 1) amounted to only about 40% of that measured during the exponential growth phase in the cyclic batch culture and continuous culture with a low initial cell concentration (Table 5). At the end of the cultivation, the MAB concentration increased steeply to 113 mg 1-1 (Fig. 21). This (10 6

.z

2~0 =

cultivation time lit]

Fig. 15. Lactate and ammonia concentrations as functions of the cultivation time during cyclic batch cultivation of the 3C2 cell line

satisfactorily (Fig. 13). For the balancing of the medium components the volume loss by sampling was taken into account.

Continuous (perfusion) cultivation of 3C2 The supplemented RPMI medium was used with 4% HS. The initial cell concentration of 3.104 cells m1-1 increased after a 20-h lag phase and 50 h exponential growth with #=0.039 h -1 to 1.1-10 6 cells m1-1. After 250 h continuous culture with a medium throughput of about 2.6 1 per day, the cultivation was changed to cyclic batch culture. During the stationary phase at a constant glucose concentration level of 0.18 g 1-~, the specific oxygen consumption rate attained values between 3 and 6 t~g

/¢".~ /

c~ I

~

\X~\ X\ ~.~ i ~k

ii

o

K

\

iii

-

o

o

o

o

m

m

•

~/~

o

_

II~

x

I ['x x

~

~-. ~ ~

iI" .~ ~ ~

~ •

~

f

I

t

I

~0

~00

~BO

I

200

cultivation time [hi

. ~

0

I

~BO

L~

3o0°

o

.

Fig. 16. Glucose, lactate and ammonia concentrations as functions of the time during continuous perfusion cultivationof the 3C2 cell line

172 '~o

ii

|

.~.

[

\

~-~--~/

.~

~,----~---... s~-~,~ ~/./x ~'~

.~-

"

~..~

I I

/ It ....

- ~ l i t - , .-~1 ~ _.,~..~

"-'~-

~._ -i~-

~-~

~'

--

I

~

~

loo

l~O

200

300d

2gO

~

cultivation time [h]

Fig. 17. Glutamine, alanine, and glycine concentrations as functions of the cultivation time during continuous perfusion cultivation of the 3C2 cell line

o~

~1

"~o '~ . . . . . .

_.."~/

~.

••

$

x'~-,,. ~

• ~1,

"

do

"~' ii1~

I

I

I

~0

tO0

1~0

.

.

~

•~ . ~ . ~ . ~

~

I

i

200

2~0

....

~

300=;

o

cultivation time [hi

~ ....

~ ' r ~

\

~ \ ,~

\

Fig. 18. Threonine, tyrosine and phenylalanine concentrations as functions of the cultivation time during continuous perfusion cultivation of the 3C2 cell line

\ ••

~ ,

~ \

\

~

~

~

X

~

o

~

~" ~,~,,.

~,

X~X~'~.X ~

~0

I SO

I I00

i .i XN.~'/

,.j,.~Z, . , ISO

/-~

_-.-

1 200

cultivation time [h]

....

~1 2S0

30~

Fig. 19. Methionine, valine, and lysine concentrations as functions of the cultivation time during continuous perfusion cultivation of the 3C2 cell line

173

,.~

~

~

-

~

,_

'~

,,\-,~\

.~

/

~~

--

.... ,

'.,

~

•~ ~ 5

!7 =

~

X'Xi'l~'l'l ..

X'X~ ~

~ i ~

I

~

SO

I I00

~# I

I~0

200

I 2S0

cultivation time [hi

Table 4. Initial concentrations (in mM l-i), concentrations (in % of the initial concentration) and specific consumption rates [in nM (10 6 cells h) -l] of amino acids and metabolites of the 3C2 cells and MAB concentration (mg 1-1) and production rates [ng (10 6 cell h) -1] during exponential growth (1), end of exponential growth (2) shortly before the feeding of fresh medium (3) and shortly after the feeding of fresh medium (4) during the stationary phase Amino acids

Culture Specific medium consumption [mM 1-1] rate (1)

Concentrations (2)

(3)

(4)

[O/o]

[O/o]

[o,~]

- 10.1 - 20.1 11.8 - 10.6 199.1 - 41.2 8.1 - 55.8 4.5 4.9 9.6 4.9 20.3 21.2 8.8

189.0 285.0 52.0 151.0 22.8 454.0 24.0 3143.0 37.0 24.5 10.4 13.6 15.9 22.3 37.9

83.4 122.0 1 8 5 . 0 238.0 50.0 41.0 69.0 119.0 0.3 14.0 372.0 391.0 13.0 18.0 3170.0 3302.0 24.0 29.0 6.9 7.4 4.8 4.8 8.7 13.6 6.3 12.6 3.1 9.5 17.7 26.5

10.68 0.21 1.18 0.13

188.0 - 563.0 - 99.8 - 9.3

75.0 3883.0 217.0 147.0

6.4 4971.0 294.0 28.5

8.0 4206.0 284.0 62.5

[rag 1-1]

[ng (106 cell h) - a] -510.5

[°R]

[°/0]

[°/0]

[nM (10 6

cells h)- 1] Asp Giu Asn Ser Gin Gly Thr Ala Thr Met Val Phe Ile Leu Lys

0.157 0.153 0.349 0.289 3.663 0.164 0.151 0.026 0.100 0.091 0.151 0.081 0.339 0.338 0.119

Other components Glucose Lactate Ammonia Acetate

MAB (IgG) 2.60

IgG, immunoglobulin G

376.0

o

30~

~

Fig. 20. Isoleucine, leucine and asparagine concentrations as functions of the cultivation time during continuous perfusion cultivation of the 3C2 cell line

was c a u s e d by retention o f M A B by the p e r f u s i o n microfiltration m e m b r a n e , since after long o p e r a t i o n it acts like an ultrafiltration m e m b r a n e , as can be seen in Fig. 22. This c h a n g e is irreversible u n d e r cultivation conditions. Except for the c o n t i n u o u s culture at high initial cell c o n c e n t r a t i o n , the specific g r o w t h rate o f the cells was fairly constant ~ = 3 . 5 to 4 . 1 0 -2 h - l ) . The low/~ value o f this run could be caused by the high t h r o u g h p u t rate (growth factors p r o d u c e d by the cells kept at a low level) or by the different qualities o f the HS or inoculum. The M A B concentrations attained in cyclic-batch o p e r a t i o n were 40% higher t h a n those in the b a t c h operation. In the perfusion operation, the M A B was enriched a n d therefore c a n n o t be c o m p a r e d with the M A B c o n c e n t r a t i o n s attained in other runs. The aim o f these investigations was the improvem e n t o f the process analysis o f h y b r i d o m a cell cultivation a n d M A B p r o d u c t i o n . On-line a n d off-line analyses allowed the m o n i t o r i n g o f several m e d i u m c o m p o n e n t s a n d the calculation o f their c o n s u m p t i o n / p r o d u c t i o n rates in different process phases. A c o m p a r i s o n o f cultivations with different sera a n d in different o p e r a t i o n m o d e s indicated that the investigated h y b r i d o m a cells with an R P M I 1640 basic med i u m s u p p l e m e n t e d with 10% FCS s h o w e d the highest p e r f o r m a n c e . Batch cultures in spinner flasks perf o r m e d m o r e efficiently t h a n in the 10-1 stirred tank investigated. I n cyclic-batch operations, higher p r o d u c t c o n c e n t r a t i o n s were attained t h a n in b a t c h cultures. D u r i n g c o n t i n u o u s perfusion M A B was enriched in the m e d i u m , because the macrofiltration p e r f u s i o n m e m b r a n e c h a n g e d its character irreversibly after longer operations a n d acted like an ultrafiltration m e m b r a n e .

1507.0 1969.0

Acknowledgements. Part of these investigations were carried out within the project 0132 D of Biotechnology Action Program of the European Community in cooperation with J. Lehmann, Bielefeld, J. M. Engasser, P. Nabet (Nancy), and J. Hache (Plaisir). The au-

174 o~