Cytotechnology 6: 121-130, 1991. 9 1991 KluwerAcademic Publishers. Printed in the Netherlands.
Selective removal of ammonia from animal cell culture broth Fidel Rey P. Nayve Jr., Masamichi Motoki, Masatoshi Matsumura and Hiroshi Kataoka Institute of Applied Biochemistry, University of Tsukuba, Tsukuba-shi, Ibaraki-ken 305, Japan Received 26 March 1991; acceptedin revisedform 7 May 1991
Key words: ammonia removal, hybridoma, HBs monoclonal antibody, zeolite
Abstract Serum-free perfusion cultures of hybridoma TO-405 cells were carried out in spinner flasks coupled with zeolite A-3 packed beads. Ammonia was selectively removed from the culture broth by passing cell free permeate from ceramic cross flow filtration, through the zeolite packed bed. Ammonia concentration in the culture broth was effectively maintained between 1 to 4 mmolfl which was below the inhibitory concentration for cell growth. Maximum cell density levels of 107 cells/ml as well as improved percentage cell viability higher than in serum-supplemented cultures were feasible in this system. The possible effects of shear stress, generated by variation of the flow rates of the broth through the ceramic filter module, on the growth of the hybridoma cells were investigated. Backwashing, by reversing the direction of the permeate, was found necessary to prolong the life of the filter. Variation of the flow rates of the broth through the ceramic module between 0.29 m/s to 0.59 m/s did not cause immediate cell damage but growth was repressed at the higher flow rate. This study also showed that glutamine appears to be one of the factors limiting the growth of the hybridoma cells.
Introduction One of the schemes proposed to achieve efficient and increased product yield from animal cell culture systems is to maximize the cell yield per unit reactor volume (Butler, 1983). This has been realized by cultivating the cells at high density. High density cultivation, however, may not only lead to high productivity but also to the potential problem of generating high concentrations of toxic metabolic end products as well. The desired compounds are not the only product of such cultivation but also included are potentially toxic metabolic end products such as ammonia. Thus, together with the need for improved volumetric
productivity, there co-exists the pressing necessity to develop a system for the reduction, if not the removal, of these toxic metabolites from animal cell culture systems. Various approaches have been proposed to solve ammonia inhibition in animal cell culture systems. Substitution of glutamine in the medium to alter cell metabolism to produce less ammonia had been proposed (Griffiths, 1973; Hosoi et al., 1988; Kurano et al., 1990), but this scheme may only be applicable to cell lines which can adapt to these substitutes. Glacken and co-workers developed mathematical relations from glutamine-ammonium culture dynamics to formulate a glutamine feeding strategy to reduce ammonia accu-
122 mulation in the culture medium (Glacken et al., 1986). The use of perfusion culture has been recommended to remove waste metabolic products from the culture system while at the same time, supplying the necessary nutrients for cell metabolism. Increased perfusion rates, however, are more often required to maintain the concentration of toxic metabolites below inhibitory levels (Butler, 1983 and 1985; Butler and Jerkins, 1989). This is not cost effective because medium components which are not yet completely exhausted are being discarded. Matsushita et al. (Abstract, JAACT 1990), proposed the use of a coupled enzyme reaction using glutamate dehydrogenase to incorporate ammonia into a-ketoglutarate and lactate dehydrogenase to convert lactate to pyruvate. Although encouraging results have been reported, much work is still left to be done to overcome the instability of the enzymes and co-enzymes involved in the coupled reactions. Membrane separation such as ultrafiltration, microfiltration and dialysis, perstraction or solvent extraction using neutral carders, adsorption and ion-exchange among others are attractive alternatives to physically remove accumulated metabolites from the culture broth. The use of some of these processes for reduction of product inhibition in microbial systems have been explored (Belfort, 1989; Le and Atkinson, 1985; McCarty, 1962). The application of these processes in reducing metabolic inhibition in animal cell cultivation systems, however, has yet to be encountered in the literature. Furthermore the specificity of these procedures must be taken into careful consideration because removal of important minute components of the medium may pose more serious problems. Mouse-mouse hybridoma TO-405 cells have been reported to be especially susceptible to ammonia inhibition in serum-free medium. The maximum density of cells cultivated in serumfree perfusion cultures was only about 70% of that in serum-supplemented cultures (Matsumura et al., 1989). Production of HBs monoclonal antibody (HBs-MAb) by the cells, however, have been also reported to be higher in serum-free
medium than in serum supplemented culture. It was thought that if ammonia could be selectively removed from the culture broth, then the maximum viable cell density in serum free cultures could be improved. In this study, the ammonia inhibition of hybridoma cell growth was investigated. A system to selectively remove ammonia from the culture broth was developed and the effect of shear stress associated with the system, on cell growth was examined.
Materials and methods Cell line and culture medium Hybridoma TO-405 cells derived from mouse myeloma NS-1 was used in the study. The hybridoma secretes monoclonal antibody against hepatitis B surface antigen (HBs-MAb). The cells were maintained routinely in E-RDF basal medium (Kyokuto Pharmaceuticals, Japan) incubated at 37~ in humidified 5% CO2 atmosphere. The basal medium was supplemented with the RD-1 I (Kyokuto Pharmaceuticals, Japan) containing the following growth factors: transferfin (10 mg/l) Na-selenite (0.0043 mg/l) ethanolamine (1.53 mg/1), insulin (9 mg/1) and sodium bicarbonate (1113 rag/l). For the perfusion cultures, the above medium was further supplemented with 1 g/1 Bovine Albumin (Sigma).
Experimental apparatus The schematic diagram of the culture system is shown in Fig. 1. The perfusion cultures were carried out in a 500 ml water jacketed spinner glass reactor (working volume: 250 ml, I.D. : 76 ram) equipped with pH, dissolved-oxygen (DO), temperature and liquid level controls. Agitation was provided by a four-bladed magnetic impeller (width: 30 mm, height 50 ram). Surface aeration and bubble-free aeration through a porous PTFE tube (TB-32, Sumitomo, Japan, 70% porosity) submerged into the broth were effected to control the DO. To retain the cells and the product in the
123
FEED
f CERAMIC
CONTINUOUS FILTER
!
BIOREACTOR
PERMEATE /
g
: ll
ZEOLITE
'
I
L,
'c,
Fig. 1. Experimental set-up for perfusion culture with ammonia removal using zeolite ion-exchange,dialysis and cross-flow ceramic filtration system.
reactor, an ultra-filter (UF) membrane (Diaflo YM10, 10000 M W cutoff, A m i c o n Corp.) was fixed at the bottom o f the reactor. The ammonia removal system consisted o f the following: a Ceramic Continuous Filter (MCF200, T o k y o Rikakikai Co., Japan) equipped with a ceramic cross-flow filter module (A1203, pore size: 0.2 ktm, thickness o f skin layer: 60 l.tm, contact area: 246 c m 2, module housing and ceramic filter clearance: 2 mm); a hollow fiber m o d u l e ( A M - N e o 1000H, effective surface area: 1 m 2, Asahi Medical Co., Japan); and 200 ml packed zeolite A-3 beads (bead size: 0.997 mm, max. pore size: 0.42 nm, Tosoh, Japan). The a m m o n i a removal system was initially primed with the cultivation medium.
Perfusion culture and ammonia removal The cultivation was started by inoculation with
approximately 1.5 x 105 cells/ml. Surface and bubble-free aeration with 5% C O 2 in air and or pure oxygen maintained the DO around 2 mg/l. The p H was controlled around 7.2 with 1 mol/1 NaOH, agitation was at 35 rpm and temperature was maintained at 37~ Perfusion o f serum-free m e d i u m at a dilution rate o f 1 day -1 and ammonia removal were started when the viable cell attained a concentration of about 1 • 106 cells/ml. A peristaltic pump controlled by the level sensor, withdrew cell-free culture broth through the ultra-filter membrane to maintain the culture volume as well as to accumulate the cells and the product HBs-MAb, in the reactor. The culture broth was filtered through a ceramic cross-flow module to obtain cell-free culture broth for ammonia removal. The broth was circulated by a peristaltic pump through the ceramic module at two linear velocities, Vs 0.59 and 0.29 m/s. Backwashing by reversing the direction
124 o f the p e r m e a t e for 3 seconds (after every 6 seconds of forward flow), was found necessary to dislodge clogging particles and to prolong the life of the filter. T h e cell-free filtrate f r o m ceramic filtration was dialyzed through a hollow fiber module at a f l o w rate o f 3.4 1/day. A m m o n i a was then selectively r e m o v e d f r o m the culture broth by circulating the dialyzed p e r m e a t e through the p a c k e d zeolite A-3 beads at a flow rate of 1.5 1/day.
Analytical methods T h e cells w e r e counted using a h e m a c y t o m e t e r and the viable cell concentration was determined using the trypan blue exclusion method. Glucose concentrations were determined by enzymatic methods using a G l u c o s e - C Test Kit (Wako, pure chemicals, Japan). A m m o n i a was m e a s u r e d by colorimetric methods using an A m m o n i a Test Kit (Wako, pure chemicals, Japan). Lactic acid concentration was m e a s u r e d b y high pressure liquid c h r o m a t o g r a p h y (Gasukuro, K o g y o , Column: Shodex I o n p a k KC-810P). Samples were also submitted for glutamine and a m i n o acid analyses (HPLC: D I O N E X , D-502) after deproteinization b y sulfosalicylic acid.
Results and discussion Selection of zeolite Certain zeolites h a v e been k n o w n to possess excellent NH4 + ion selectivity. This has led to the c o m m e r c i a l application o f zeolites in the r e m o v a l of NH4 + f r o m wastewater (Sherman, 1978). For application o f zeolites in animal cell culture system, the selection o f suitable type of zeolite that has a strong affinity for NH4 + but not for other c o m p o n e n t s of the m e d i u m must be considered. The selectivity of 3 synthetic and 2 natural zeolites were evaluated. All exchangeable cations in the zeolite were replaced with sodium ions b y soaking the beads in 1 mg/1 N a O H for at least two hours at r o o m temperature. T h e N a - f o r m zeolite beads were then incubated in E - R D F m e d i u m supplemented with 10% FCS and 1.46 mmol/1 a m m o n i u m chloride for 14 h at 37~ The equilib r i u m concentrations o f a m m o n i a and amino acids in the m e d i u m are listed in T a b l e 1 as percentages of their original compositions. The ion-exchange capacity for a m m o n i a were almost similar h o w e v e r , the uptake o f s o m e a m i n o acids particularly lysine and arginine varied significantly between zeolites. Since the molecular
Table 1. Removal of ammonia and amino acids from E-RDF medium by different types of zeolites (percentage of original conc.) (NB. Variation of Lysine and Arginine with different zeolites) Zeolite type origin
Initial concentration (m-mol/m3)
Zeolite A-3 A-type Zeolite synthetic
Zeolite F-9 X-type Zeolite synthetic
Silica-aluminium ZCP-50 Y-type Zeolite synthetic
clinoptilolite ditto natural
mordenite ditto natural
NH3 ASP GLU CYS PHE GLN TYR MET ILE LEU ALA HIS LYS ARG
1462 73.48 108.0 93.55 109.7 1522 113.8 70.89 281.8 296.1 65.04 85.62 272.0 646.4
-51.76 --0.41 +1.56 -0.43 -0.92 -6.34 +2.19 +1.68 +0.99 +1.84 +0.91 -12.69 -8.87 -1.58
-68.69 -0.66 +4.69 -10.61 +0.04 -10.43 +1.96 -8.60 +2.19 +1.93 +0.76 -24.32 -68.67 -28.56
-55.48 +3.13 +0.19 -2.37 -0.80 -10.78 -2.16 -4.52 -0.81 -0.01 +2.34 -1.64 -73.23 --61.38
-49.19 -4.33 +2.54 -10.48 +6.62 -16.31 -1.26 -11.87 - 1.94 -3.56 -5.09 -22.40 -20.40 -13.79
-58.58 -5.46 +4.41 -15.60 -5.88 -19.26 -3.03 -12.48 -2.72 -3.95 -12.43 -20.29 -35.69 -32.77
125 weight of amino acids are about 10 times than that of ammonia, the high selectivity of zeolite A-3 seems to be due to the molecular sieving effect. For this study, therefore, zeolite A-3 was selected.
Ion-exchange capacity of Zeolite A-3 Ammonium ion-exchange isotherms of the zeolite A-3 beads were measured in aqueous solution of ternary system of Na + - K § - NH4 + and E-RDF medium supplemented with 1 g/l BSA. E-RDF medium contains 130 mmol/1 Na + and 5 mmol/1 K § therefore, the zeolite beads were pretreated with a mixture of NaC1 (130 mmol/1) and KC1 (5 mmol/1). The pretreated zeolite beads were equilibrated with temary cation solutions containing different ammonia concentrations, but their total ion concentration and Na + and K + ratio were fixed at 135 mmol/l and 130:5 respectively. Different amounts of zeolite beads were also equilibrated in 10 ml solution of E-RDF medium and NH4C1 (130 mmol/1) mixed at 9:1 ratio. The results of these experiments are shown in Fig. 2. The isotherms were straight and accorded to Henry's law in the experimental region. This 30 25
// // / / , , ~ / s ~ J // J
Co = 185 mol/m~ tlo = 3.7~ mol/kg H = 9..77x10-z m~/kg ,//
"~20 E 15
~_1o 0
0
I
I
2
4
I
C
I
6 8 ( m o l / m 3)
I
I
10
12
Fig. 2. Ammonium ion-exchange isotherms of zeolite A-3 beads in NH4+-Na+-K+ ternary aqueous solution and E-RDF + 0.1% BSA medium (D NH4+-Na+-K+ soln.; O E-RDF + 0.1% BSA medium).
suggests that there is no large difference in the selectivity between Na § K + and N H 4 4 . From the equation of the straight lines, the equilibrium constant H and the exchange capacity q0 in the ternary solutions were computed to be 2.77 • 10-2 m3/kg and 3.74 mol/kg, while H for the E-RDF medium was slightly lower at 2.50 x 10-2 m3/kg. E-RDF medium contains other metallic cations such as Mg 2+, Ca 2+, Zn 2+ and Cu 2+ hence, the slight drop in the H value. The effect of the other cations in the medium on H were not considered so significant. The high values of H and ion-exchange capacity obtained in this experiment were considered to justify the use of zeolite A-3 for the selective removal of ammonia from animal cell culture broth.
Characteristics of hybridoma T0-405 cells HBs monoclonal antibody production has been previously demonstrated to be growth associated and proportional to the viable cell density. Although the maximum viable cell densities is lower in serum-free cultures, HBs-MAb production is higher compared to that in serum cultures (Matsumura et al., 1989). The sensitivity of TO-405 cells to ammonia inhibition was investigated by carrying out perfusion cultures in E-RDF medium supplemented with 10% FCS under different DO levels. The cell viability appeared to be inhibited when ammonia accumulated to about 5 mmol/1 regardless of the DO levels (Fig. 3). Previous results also showed that although the final ammonia concentration were similar in serum supplemented and serum-free cultures, the rate of accumulation was faster with the latter (Misato et al., Abstracts of the Annual Meeting of the Soc. of Chem. Engrs. Japan, 1989, p. 565).
Perfusion culture with ammonia removal Perfusion cultures without ammonia removal were carried out as a control experiment. The results are shown in Fig. 4. The cells continued to
126 100
o
o
o
o
80
~
~ 60
o
23 .~ 4 0
eO
20
0
I
I
I
I
I
I
I
1
2
3
4
5
6
7
8
Ammonia conc. (mmol/I)
Fig. 3. Effect of ammonia concentration on the viability of TO-405 hybridoma during perfusion cultures with E-RDF + 10% Fetal Calf Serum. The points represents the values under different dissolve oxygen conditions (C) 0.27 ppm; ~ 0.71 ppm; 3.57 ppm).
grow, h o w e v e r , the viable cell density did not i m p r o v e further f r o m about 6.6 x 106 cells/ml. T h e percent viability then continued to decrease. This was considered to be caused b y a m m o n i a toxicity. A m m o n i a concentration in the m e d i u m a c c u m u l a t e d to inhibitory level and r e m a i n e d a b o v e 5 mmol/1 in spite o f perfusion. This clearly shows that the rate o f perfusion at 1 culture v o l u m e per day was not sufficient to prevent the a c c u m u l a t i o n o f a m m o n i a to toxic levels in this culture. A higher perfusion rate m a y be needed but it m i g h t p r o v e to be costly.
Perfusion culture with a m m o n i a r e m o v a l was, therefore, carried out. T h e cell growth, however, was c o m p l e t e l y repressed in perfusion culture where the linear velocity o f the broth circulation through the ceramic filter was as high as 0.59 rn/s (Fig. 5). This was considered to be due to the excessive shear stress generated by the circulation o f the broth through the ceramic filter. This is described further in the next section. T h e linear velocity o f broth circulation was then reduced to 0.29 rn/s. The results are shown in Fig. 6. A m a x i m u m viable cell density o f 1.2 x 107 cell/ml was attained. The viability was also maintained a b o v e 90% throughout the cultivation. This was attributed to the r e m o v a l of growth inhibition or toxicity due to accumulation o f ammonia. A m m o n i a was effectively maintained b e l o w the inhibitory level. Glucose concentration in the broth r e m a i n e d high. Glutamine h o w e v e r was almost depleted specifically during the later part o f the cultivation. Glutamine concentration in the m e d i u m was 0.03 m m o l / l and below, and this m a y be the reason w h y the cell growth s e e m e d to slow d o w n and enter a stationary phase. Glutamine has been recognized to be one of the highly utilized m e d i u m c o m p o n e n t in cell cultures. G r o w t h limitation due to an almost depleted concentration of glutamine in the culture m e d i u m has been o b s e r v e d in a wide variety of
t~
s ,,\
mTA~U-
~..-~
2o~
8
u
20
d
.-
E o
d co= 10 'm
g GLUTAMINE
.~/ 20 20
40
60
80 - 100
120
1210
160
180- 200 ~ 220
s~
- - - AMMONIA 40
60
80
100
120
140
160
180
o
Time (h)
Time (h)
Fig. 5. Perfusion culture of TO-405 hybridoma cells coupled Fig. 4. Perfusion culture of TO-405 hybridoma cells with
E-RDF + 1 g/l BSA. The arrow indicates the time when perfusion was started.
with the ammonia removal system. Perfusion culture and ammonia removal were started simultaneously as indicated by the arrow. The circulation of the culture broth through the ceramic module was 0.59 m/s.
127 cell lines (Butler, 1985; Griffiths and Pirt, 1967; McCarty, 1962; Roberts, 1976). Supplementation of glutamine in these cultures has resulted in increased population density, a shorter lag phase, and increased secretion of monoclonal antibody (Geaugey et al., 1989; Griffiths and Pirt, 1967; Griffiths, 1970). One of the apprehensions, about increasing the concentration of glutamine in cell culture medium is the possibility of generating too much ammonia and the consequent accumulation to inhibitory or toxic levels. It must be considered that the enzymatic deamination of one mole of glutamine to glutamic acid and the subsequent deamination of glutamic acid can release two moles of ammonium (Doyle, 1990). It is also well known that glutamine is quite unstable in cell culture media at elevated temperatures. The spontaneous decomposition of glutamine, catalyzed by phosphate and bicarbonate ions at 37~ to pyrrolidone carboxylic acid can further release ammonia (Griffiths and Pirt, 1967). In this study, however, it has been demonstrated that the ammonia in the medium can be removed and the concentration can be maintained below inhibitory or toxic levels by using the ammonia removal system proposed. It may thus be possible to take full advantage of the potential benefits resulting from supplementation of cultures with the optimum glutamine concentration without fear of ammonia inhibition.
F•
TOTALCELL
"-O d
PERFUSION C~ILTURE
.m
"
Selective removal of ammonia from animal cell culture broth Fidel Rey P. Nayve Jr., Masamichi Motoki, Masatoshi Matsumura and Hiroshi Kataoka Institute of Applied Biochemistry, University of Tsukuba, Tsukuba-shi, Ibaraki-ken 305, Japan Received 26 March 1991; acceptedin revisedform 7 May 1991
Key words: ammonia removal, hybridoma, HBs monoclonal antibody, zeolite
Abstract Serum-free perfusion cultures of hybridoma TO-405 cells were carried out in spinner flasks coupled with zeolite A-3 packed beads. Ammonia was selectively removed from the culture broth by passing cell free permeate from ceramic cross flow filtration, through the zeolite packed bed. Ammonia concentration in the culture broth was effectively maintained between 1 to 4 mmolfl which was below the inhibitory concentration for cell growth. Maximum cell density levels of 107 cells/ml as well as improved percentage cell viability higher than in serum-supplemented cultures were feasible in this system. The possible effects of shear stress, generated by variation of the flow rates of the broth through the ceramic filter module, on the growth of the hybridoma cells were investigated. Backwashing, by reversing the direction of the permeate, was found necessary to prolong the life of the filter. Variation of the flow rates of the broth through the ceramic module between 0.29 m/s to 0.59 m/s did not cause immediate cell damage but growth was repressed at the higher flow rate. This study also showed that glutamine appears to be one of the factors limiting the growth of the hybridoma cells.
Introduction One of the schemes proposed to achieve efficient and increased product yield from animal cell culture systems is to maximize the cell yield per unit reactor volume (Butler, 1983). This has been realized by cultivating the cells at high density. High density cultivation, however, may not only lead to high productivity but also to the potential problem of generating high concentrations of toxic metabolic end products as well. The desired compounds are not the only product of such cultivation but also included are potentially toxic metabolic end products such as ammonia. Thus, together with the need for improved volumetric
productivity, there co-exists the pressing necessity to develop a system for the reduction, if not the removal, of these toxic metabolites from animal cell culture systems. Various approaches have been proposed to solve ammonia inhibition in animal cell culture systems. Substitution of glutamine in the medium to alter cell metabolism to produce less ammonia had been proposed (Griffiths, 1973; Hosoi et al., 1988; Kurano et al., 1990), but this scheme may only be applicable to cell lines which can adapt to these substitutes. Glacken and co-workers developed mathematical relations from glutamine-ammonium culture dynamics to formulate a glutamine feeding strategy to reduce ammonia accu-
122 mulation in the culture medium (Glacken et al., 1986). The use of perfusion culture has been recommended to remove waste metabolic products from the culture system while at the same time, supplying the necessary nutrients for cell metabolism. Increased perfusion rates, however, are more often required to maintain the concentration of toxic metabolites below inhibitory levels (Butler, 1983 and 1985; Butler and Jerkins, 1989). This is not cost effective because medium components which are not yet completely exhausted are being discarded. Matsushita et al. (Abstract, JAACT 1990), proposed the use of a coupled enzyme reaction using glutamate dehydrogenase to incorporate ammonia into a-ketoglutarate and lactate dehydrogenase to convert lactate to pyruvate. Although encouraging results have been reported, much work is still left to be done to overcome the instability of the enzymes and co-enzymes involved in the coupled reactions. Membrane separation such as ultrafiltration, microfiltration and dialysis, perstraction or solvent extraction using neutral carders, adsorption and ion-exchange among others are attractive alternatives to physically remove accumulated metabolites from the culture broth. The use of some of these processes for reduction of product inhibition in microbial systems have been explored (Belfort, 1989; Le and Atkinson, 1985; McCarty, 1962). The application of these processes in reducing metabolic inhibition in animal cell cultivation systems, however, has yet to be encountered in the literature. Furthermore the specificity of these procedures must be taken into careful consideration because removal of important minute components of the medium may pose more serious problems. Mouse-mouse hybridoma TO-405 cells have been reported to be especially susceptible to ammonia inhibition in serum-free medium. The maximum density of cells cultivated in serumfree perfusion cultures was only about 70% of that in serum-supplemented cultures (Matsumura et al., 1989). Production of HBs monoclonal antibody (HBs-MAb) by the cells, however, have been also reported to be higher in serum-free
medium than in serum supplemented culture. It was thought that if ammonia could be selectively removed from the culture broth, then the maximum viable cell density in serum free cultures could be improved. In this study, the ammonia inhibition of hybridoma cell growth was investigated. A system to selectively remove ammonia from the culture broth was developed and the effect of shear stress associated with the system, on cell growth was examined.
Materials and methods Cell line and culture medium Hybridoma TO-405 cells derived from mouse myeloma NS-1 was used in the study. The hybridoma secretes monoclonal antibody against hepatitis B surface antigen (HBs-MAb). The cells were maintained routinely in E-RDF basal medium (Kyokuto Pharmaceuticals, Japan) incubated at 37~ in humidified 5% CO2 atmosphere. The basal medium was supplemented with the RD-1 I (Kyokuto Pharmaceuticals, Japan) containing the following growth factors: transferfin (10 mg/l) Na-selenite (0.0043 mg/l) ethanolamine (1.53 mg/1), insulin (9 mg/1) and sodium bicarbonate (1113 rag/l). For the perfusion cultures, the above medium was further supplemented with 1 g/1 Bovine Albumin (Sigma).
Experimental apparatus The schematic diagram of the culture system is shown in Fig. 1. The perfusion cultures were carried out in a 500 ml water jacketed spinner glass reactor (working volume: 250 ml, I.D. : 76 ram) equipped with pH, dissolved-oxygen (DO), temperature and liquid level controls. Agitation was provided by a four-bladed magnetic impeller (width: 30 mm, height 50 ram). Surface aeration and bubble-free aeration through a porous PTFE tube (TB-32, Sumitomo, Japan, 70% porosity) submerged into the broth were effected to control the DO. To retain the cells and the product in the
123
FEED
f CERAMIC
CONTINUOUS FILTER
!
BIOREACTOR
PERMEATE /
g
: ll
ZEOLITE
'
I
L,
'c,
Fig. 1. Experimental set-up for perfusion culture with ammonia removal using zeolite ion-exchange,dialysis and cross-flow ceramic filtration system.
reactor, an ultra-filter (UF) membrane (Diaflo YM10, 10000 M W cutoff, A m i c o n Corp.) was fixed at the bottom o f the reactor. The ammonia removal system consisted o f the following: a Ceramic Continuous Filter (MCF200, T o k y o Rikakikai Co., Japan) equipped with a ceramic cross-flow filter module (A1203, pore size: 0.2 ktm, thickness o f skin layer: 60 l.tm, contact area: 246 c m 2, module housing and ceramic filter clearance: 2 mm); a hollow fiber m o d u l e ( A M - N e o 1000H, effective surface area: 1 m 2, Asahi Medical Co., Japan); and 200 ml packed zeolite A-3 beads (bead size: 0.997 mm, max. pore size: 0.42 nm, Tosoh, Japan). The a m m o n i a removal system was initially primed with the cultivation medium.
Perfusion culture and ammonia removal The cultivation was started by inoculation with
approximately 1.5 x 105 cells/ml. Surface and bubble-free aeration with 5% C O 2 in air and or pure oxygen maintained the DO around 2 mg/l. The p H was controlled around 7.2 with 1 mol/1 NaOH, agitation was at 35 rpm and temperature was maintained at 37~ Perfusion o f serum-free m e d i u m at a dilution rate o f 1 day -1 and ammonia removal were started when the viable cell attained a concentration of about 1 • 106 cells/ml. A peristaltic pump controlled by the level sensor, withdrew cell-free culture broth through the ultra-filter membrane to maintain the culture volume as well as to accumulate the cells and the product HBs-MAb, in the reactor. The culture broth was filtered through a ceramic cross-flow module to obtain cell-free culture broth for ammonia removal. The broth was circulated by a peristaltic pump through the ceramic module at two linear velocities, Vs 0.59 and 0.29 m/s. Backwashing by reversing the direction
124 o f the p e r m e a t e for 3 seconds (after every 6 seconds of forward flow), was found necessary to dislodge clogging particles and to prolong the life of the filter. T h e cell-free filtrate f r o m ceramic filtration was dialyzed through a hollow fiber module at a f l o w rate o f 3.4 1/day. A m m o n i a was then selectively r e m o v e d f r o m the culture broth by circulating the dialyzed p e r m e a t e through the p a c k e d zeolite A-3 beads at a flow rate of 1.5 1/day.
Analytical methods T h e cells w e r e counted using a h e m a c y t o m e t e r and the viable cell concentration was determined using the trypan blue exclusion method. Glucose concentrations were determined by enzymatic methods using a G l u c o s e - C Test Kit (Wako, pure chemicals, Japan). A m m o n i a was m e a s u r e d by colorimetric methods using an A m m o n i a Test Kit (Wako, pure chemicals, Japan). Lactic acid concentration was m e a s u r e d b y high pressure liquid c h r o m a t o g r a p h y (Gasukuro, K o g y o , Column: Shodex I o n p a k KC-810P). Samples were also submitted for glutamine and a m i n o acid analyses (HPLC: D I O N E X , D-502) after deproteinization b y sulfosalicylic acid.
Results and discussion Selection of zeolite Certain zeolites h a v e been k n o w n to possess excellent NH4 + ion selectivity. This has led to the c o m m e r c i a l application o f zeolites in the r e m o v a l of NH4 + f r o m wastewater (Sherman, 1978). For application o f zeolites in animal cell culture system, the selection o f suitable type of zeolite that has a strong affinity for NH4 + but not for other c o m p o n e n t s of the m e d i u m must be considered. The selectivity of 3 synthetic and 2 natural zeolites were evaluated. All exchangeable cations in the zeolite were replaced with sodium ions b y soaking the beads in 1 mg/1 N a O H for at least two hours at r o o m temperature. T h e N a - f o r m zeolite beads were then incubated in E - R D F m e d i u m supplemented with 10% FCS and 1.46 mmol/1 a m m o n i u m chloride for 14 h at 37~ The equilib r i u m concentrations o f a m m o n i a and amino acids in the m e d i u m are listed in T a b l e 1 as percentages of their original compositions. The ion-exchange capacity for a m m o n i a were almost similar h o w e v e r , the uptake o f s o m e a m i n o acids particularly lysine and arginine varied significantly between zeolites. Since the molecular
Table 1. Removal of ammonia and amino acids from E-RDF medium by different types of zeolites (percentage of original conc.) (NB. Variation of Lysine and Arginine with different zeolites) Zeolite type origin
Initial concentration (m-mol/m3)
Zeolite A-3 A-type Zeolite synthetic
Zeolite F-9 X-type Zeolite synthetic
Silica-aluminium ZCP-50 Y-type Zeolite synthetic
clinoptilolite ditto natural
mordenite ditto natural
NH3 ASP GLU CYS PHE GLN TYR MET ILE LEU ALA HIS LYS ARG
1462 73.48 108.0 93.55 109.7 1522 113.8 70.89 281.8 296.1 65.04 85.62 272.0 646.4
-51.76 --0.41 +1.56 -0.43 -0.92 -6.34 +2.19 +1.68 +0.99 +1.84 +0.91 -12.69 -8.87 -1.58
-68.69 -0.66 +4.69 -10.61 +0.04 -10.43 +1.96 -8.60 +2.19 +1.93 +0.76 -24.32 -68.67 -28.56
-55.48 +3.13 +0.19 -2.37 -0.80 -10.78 -2.16 -4.52 -0.81 -0.01 +2.34 -1.64 -73.23 --61.38
-49.19 -4.33 +2.54 -10.48 +6.62 -16.31 -1.26 -11.87 - 1.94 -3.56 -5.09 -22.40 -20.40 -13.79
-58.58 -5.46 +4.41 -15.60 -5.88 -19.26 -3.03 -12.48 -2.72 -3.95 -12.43 -20.29 -35.69 -32.77
125 weight of amino acids are about 10 times than that of ammonia, the high selectivity of zeolite A-3 seems to be due to the molecular sieving effect. For this study, therefore, zeolite A-3 was selected.
Ion-exchange capacity of Zeolite A-3 Ammonium ion-exchange isotherms of the zeolite A-3 beads were measured in aqueous solution of ternary system of Na + - K § - NH4 + and E-RDF medium supplemented with 1 g/l BSA. E-RDF medium contains 130 mmol/1 Na + and 5 mmol/1 K § therefore, the zeolite beads were pretreated with a mixture of NaC1 (130 mmol/1) and KC1 (5 mmol/1). The pretreated zeolite beads were equilibrated with temary cation solutions containing different ammonia concentrations, but their total ion concentration and Na + and K + ratio were fixed at 135 mmol/l and 130:5 respectively. Different amounts of zeolite beads were also equilibrated in 10 ml solution of E-RDF medium and NH4C1 (130 mmol/1) mixed at 9:1 ratio. The results of these experiments are shown in Fig. 2. The isotherms were straight and accorded to Henry's law in the experimental region. This 30 25
// // / / , , ~ / s ~ J // J
Co = 185 mol/m~ tlo = 3.7~ mol/kg H = 9..77x10-z m~/kg ,//
"~20 E 15
~_1o 0
0
I
I
2
4
I
C
I
6 8 ( m o l / m 3)
I
I
10
12
Fig. 2. Ammonium ion-exchange isotherms of zeolite A-3 beads in NH4+-Na+-K+ ternary aqueous solution and E-RDF + 0.1% BSA medium (D NH4+-Na+-K+ soln.; O E-RDF + 0.1% BSA medium).
suggests that there is no large difference in the selectivity between Na § K + and N H 4 4 . From the equation of the straight lines, the equilibrium constant H and the exchange capacity q0 in the ternary solutions were computed to be 2.77 • 10-2 m3/kg and 3.74 mol/kg, while H for the E-RDF medium was slightly lower at 2.50 x 10-2 m3/kg. E-RDF medium contains other metallic cations such as Mg 2+, Ca 2+, Zn 2+ and Cu 2+ hence, the slight drop in the H value. The effect of the other cations in the medium on H were not considered so significant. The high values of H and ion-exchange capacity obtained in this experiment were considered to justify the use of zeolite A-3 for the selective removal of ammonia from animal cell culture broth.
Characteristics of hybridoma T0-405 cells HBs monoclonal antibody production has been previously demonstrated to be growth associated and proportional to the viable cell density. Although the maximum viable cell densities is lower in serum-free cultures, HBs-MAb production is higher compared to that in serum cultures (Matsumura et al., 1989). The sensitivity of TO-405 cells to ammonia inhibition was investigated by carrying out perfusion cultures in E-RDF medium supplemented with 10% FCS under different DO levels. The cell viability appeared to be inhibited when ammonia accumulated to about 5 mmol/1 regardless of the DO levels (Fig. 3). Previous results also showed that although the final ammonia concentration were similar in serum supplemented and serum-free cultures, the rate of accumulation was faster with the latter (Misato et al., Abstracts of the Annual Meeting of the Soc. of Chem. Engrs. Japan, 1989, p. 565).
Perfusion culture with ammonia removal Perfusion cultures without ammonia removal were carried out as a control experiment. The results are shown in Fig. 4. The cells continued to
126 100
o
o
o
o
80
~
~ 60
o
23 .~ 4 0
eO
20
0
I
I
I
I
I
I
I
1
2
3
4
5
6
7
8
Ammonia conc. (mmol/I)
Fig. 3. Effect of ammonia concentration on the viability of TO-405 hybridoma during perfusion cultures with E-RDF + 10% Fetal Calf Serum. The points represents the values under different dissolve oxygen conditions (C) 0.27 ppm; ~ 0.71 ppm; 3.57 ppm).
grow, h o w e v e r , the viable cell density did not i m p r o v e further f r o m about 6.6 x 106 cells/ml. T h e percent viability then continued to decrease. This was considered to be caused b y a m m o n i a toxicity. A m m o n i a concentration in the m e d i u m a c c u m u l a t e d to inhibitory level and r e m a i n e d a b o v e 5 mmol/1 in spite o f perfusion. This clearly shows that the rate o f perfusion at 1 culture v o l u m e per day was not sufficient to prevent the a c c u m u l a t i o n o f a m m o n i a to toxic levels in this culture. A higher perfusion rate m a y be needed but it m i g h t p r o v e to be costly.
Perfusion culture with a m m o n i a r e m o v a l was, therefore, carried out. T h e cell growth, however, was c o m p l e t e l y repressed in perfusion culture where the linear velocity o f the broth circulation through the ceramic filter was as high as 0.59 rn/s (Fig. 5). This was considered to be due to the excessive shear stress generated by the circulation o f the broth through the ceramic filter. This is described further in the next section. T h e linear velocity o f broth circulation was then reduced to 0.29 rn/s. The results are shown in Fig. 6. A m a x i m u m viable cell density o f 1.2 x 107 cell/ml was attained. The viability was also maintained a b o v e 90% throughout the cultivation. This was attributed to the r e m o v a l of growth inhibition or toxicity due to accumulation o f ammonia. A m m o n i a was effectively maintained b e l o w the inhibitory level. Glucose concentration in the broth r e m a i n e d high. Glutamine h o w e v e r was almost depleted specifically during the later part o f the cultivation. Glutamine concentration in the m e d i u m was 0.03 m m o l / l and below, and this m a y be the reason w h y the cell growth s e e m e d to slow d o w n and enter a stationary phase. Glutamine has been recognized to be one of the highly utilized m e d i u m c o m p o n e n t in cell cultures. G r o w t h limitation due to an almost depleted concentration of glutamine in the culture m e d i u m has been o b s e r v e d in a wide variety of
t~
s ,,\
mTA~U-
~..-~
2o~
8
u
20
d
.-
E o
d co= 10 'm
g GLUTAMINE
.~/ 20 20
40
60
80 - 100
120
1210
160
180- 200 ~ 220
s~
- - - AMMONIA 40
60
80
100
120
140
160
180
o
Time (h)
Time (h)
Fig. 5. Perfusion culture of TO-405 hybridoma cells coupled Fig. 4. Perfusion culture of TO-405 hybridoma cells with
E-RDF + 1 g/l BSA. The arrow indicates the time when perfusion was started.
with the ammonia removal system. Perfusion culture and ammonia removal were started simultaneously as indicated by the arrow. The circulation of the culture broth through the ceramic module was 0.59 m/s.
127 cell lines (Butler, 1985; Griffiths and Pirt, 1967; McCarty, 1962; Roberts, 1976). Supplementation of glutamine in these cultures has resulted in increased population density, a shorter lag phase, and increased secretion of monoclonal antibody (Geaugey et al., 1989; Griffiths and Pirt, 1967; Griffiths, 1970). One of the apprehensions, about increasing the concentration of glutamine in cell culture medium is the possibility of generating too much ammonia and the consequent accumulation to inhibitory or toxic levels. It must be considered that the enzymatic deamination of one mole of glutamine to glutamic acid and the subsequent deamination of glutamic acid can release two moles of ammonium (Doyle, 1990). It is also well known that glutamine is quite unstable in cell culture media at elevated temperatures. The spontaneous decomposition of glutamine, catalyzed by phosphate and bicarbonate ions at 37~ to pyrrolidone carboxylic acid can further release ammonia (Griffiths and Pirt, 1967). In this study, however, it has been demonstrated that the ammonia in the medium can be removed and the concentration can be maintained below inhibitory or toxic levels by using the ammonia removal system proposed. It may thus be possible to take full advantage of the potential benefits resulting from supplementation of cultures with the optimum glutamine concentration without fear of ammonia inhibition.
F•
TOTALCELL
"-O d
PERFUSION C~ILTURE
.m
"
