A comparison of intensive cell culture bioreactors ... · mechanism of cell death during the culture of many industrially important mammalian cell lines. The levels of apoptosis in

Journal of Biotechnology 62 (1998) 195–207

A comparison of intensive cell culture bioreactors operatingwith Hybridomas modified for inhibited apoptotic response.

H. Bierau, A. Perani, M. Al-Rubeai, A.N. Emery *

Centre for Bioprocess Engineering, School of Chemical Engineering, Uni6ersity of Birmingham, Edgbaston,Birmingham B15 2TT, UK

Received 10 December 1997; received in revised form 14 April 1998; accepted 27 April 1998

Abstract

It is demonstrated, using two different perfusion reaction systems, that hybridoma modified by inhibiting theirapoptotic response can give improved process performance in terms of cell number and viability in intensive cellculture. Two cell perfusion systems, one using a spin filter and the other an ultrasonic filter, are compared using twocell lines. One cell line is transfected with the bcl-2 gene (TB/C3 bcl-2) which encodes the ‘anti-apoptotic’ human bcl-2protein and the other cell line (TB/C3 pEF) with a negative transfection vector. Both reactor systems give similarretention performance for both cell lines. Bcl-2 transfected cells reach higher cell densities than the control cell line,and the percentage of apoptotic cells is clearly lower than with pEF cells. The maximum cell numbers of the bcl-2 cellline are 1.21×107 ml−1 in the ultrasonic filter culture and 1.58×107 ml−1 in the spin filter culture, respectively.Using the pEF cell line the maximum cell number reaches 6.0×106 ml−1 with ultrasonic retention and 5.9×106

ml−1 in the spin filter. The use of ultrasound in this cell retention system has no apparent influence on cell growth,productivity or viability. Selective retention of viable cells is detectable but the effect of removing non-viable cells isnegligible. © 1998 Elsevier Science B.V. All rights reserved.

Keywords: Hybridoma; Perfusion culture; Apoptosis; Bcl-2; Spin filter; Ultrasonic filter

1. Introduction1.1. Process intensification and cell death

A major problem for processes based on the

culture of animal cells is that, compared to bacte-rial cultures, specific production rates and there-fore yields are low and large numbers of cells arerequired. The most obvious way to intensify sucha process is to increase the concentration of (vi-able, producing) cells. This can be achieved byretaining cells in the vessel whilst removing spentmedium which contains the product and supply-

* Corresponding author. Tel.: +44 121 4145281; fax: 44121 4145324; e-mail: [email protected]

0168-1656/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved.

PII S0168-1656(98)00064-9

H. Bierau et al. / Journal of Biotechnology 62 (1998) 195–207196

ing fresh medium continuously. Various systemsfor cell retention have been developed, includingmembrane filtration systems, dialysis reactors,devices using gravitational settling, the spin filterand the ultrasonic filter (see review by Griffith,1990). These systems allow one to reach manytimes the cell density of batch cultures. Cellsmay be totally retained or may be bled from thesystem.

However, in many intensified processes, theculture conditions which then result due to highcell densities and insufficient mixing may lead tosignificant cell death. In particular, programmedcell death, also known as apoptosis (Kerr et al.,1972), is triggered in many cell lines by,amongst other factors, deprivation of nutrientsand–or accumulation of cytotoxic products(Singh et al., 1994, 1996a,b). Apoptosis is a pro-cess whereby developmental or environmentalstimuli activate a genetic programme to imple-ment a specific series of events that culminate inthe death and efficient disposal of a cell. In con-trast to the passive process of necrosis, apopto-sis involves the full participation of the cell inthe process of death. Experiments have increas-ingly revealed that apoptosis is the primarymechanism of cell death during the culture ofmany industrially important mammalian celllines. The levels of apoptosis in hybridoma cul-tures have been found to be up to 90% of totaldead cells at the end of batch cultures. Themost effective inducers of apoptosis have beenfound to be glutamine limitation, serum limita-tion, glucose limitation, and ammonia toxicity(Singh et al., 1994), any or all of which may beencountered in a bioreactor. Other environmen-tal signals that lead to the induction of anapoptotic response have included oxygen limita-tion and hydrodynamic stress (Al-Rubeai et al.,1995; Singh et al., 1996a). The predominatingmode of cell death is of course important in theselection of a retention system for intensive cellculture. During cell death, apoptosis leads toshrinkage of the cells while necrosis leads toswelling, but in the break-up of cells after death,the products of necrosis will be more diffusethan the defined bodies resulting from apoptotic

death (Singh et al., 1996b). Retention systemsbased on size selection, e.g. by gravity or filterpore size, will therefore react differently to thetwo modes of death.

In intensive systems, where cells are retainedover long time-periods in a potentially hostileenvironment, there is a need for the improve-ment in robustness of the cells to avoid earlycell death. A promising approach for the pro-duction of more robust cells is to manipulatetheir responses to the external environment.Such cells can be cultured in simplified mediaand should maintain low death rates under thestressful conditions encountered in bioreactors.It has been demonstrated previously (Itoh et al.,1995; Singh et al., 1994, 1995, 1996a,b; Simpsonet al., 1997) that the transfection of hybridomaand plasmacytoma cell lines with a gene thatinhibits the apoptotic response can markedly en-hance survivability to reaction conditions, in-cluding nutrient deprivation and under orover-supply of oxygen, which adversely affectthe control cell line.

One of the most widely studied apoptosis sup-pressor genes is the proto-oncogene bcl-2 (Tsuji-moto et al., 1985), a member of a group ofrelated genes that share extensive sequence ho-mology. Bcl-2 encodes a 27-kDa transmembraneprotein which is associated with the endoplasmicreticulum, nuclear envelope, and the inner mito-chondrial membrane. In vivo, bcl-2 expression isconfined to long-lived or proliferating adult cellpopulations such as memory B-cells, peripheralT-cells, haemopoietic stem cells, skin and intesti-nal epithelial stem cells, and post mitotic neu-rons.

In cultured cells the induction of apoptosiscan be suppressed by transfection with bcl-2. Itsoverexpression can protect the cell against awide range of potentially damaging agents, cyto-toxic drugs, growth factor withdrawal and de-pletion of nutrients. Itoh et al. (1995) andSimpson et al. (1997) used bcl-2 transfection ofhybridomas and the overexpression of the bcl-2gene significantly extended the cells viability bysuppression of apoptosis in conditions of hy-poxia, hyperoxia, glutamine deprivation, glucosedeprivation and serum limitation.

H. Bierau et al. / Journal of Biotechnology 62 (1998) 195–207 197

Fig. 1. Schematic configuration of the bioreactor: (A) spin filter, (B) ultrasonic filter (silicone tube not displayed).

1.2. Ultrasonic separation

A recent development in retention systems hasbeen the ultrasound filter. This system uses highfrequency resonant ultrasonic waves to separatecells instead of a physical mesh or membrane(Trampler et al., 1994; Phylis et al., 1995). Acous-tic particle separation technology utilises the prin-ciple of aggregation in a standing acoustic field toseparate suspended cells from liquid medium.Cells are exposed to an acoustic resonance field atapproximately 2 MHz, causing them to migratefirst into the antinode planes and then into theregions of highest acoustic amplitude of theseplanes where they form loose aggregates withinthe resonator. Periodic shut-off of the resonancefield allows the aggregates to settle and return to

the bioreactor. The maximum flow rate dependson cell line and concentration. Lower flow ratesshould generally result in increased separationefficiency.

The basic system used in this work (the modelADI 1015 Bio Sep from Applikon) consists of acontrol cabinet with an ultrasonic resonance gen-erator and a control unit which is connected to a10 ml resonance chamber. The in situ sterilizableresonance chamber is installed into a standard 12mm port in the bioreactor headplate. Two peri-staltic pumps are required for recirculating cellsuspension and harvesting.

Fig. 1 shows a typical configuration of the cellseparation system with the resonator installed intothe bioreactor headplate. The recirculation pumppumps cell suspension from the bioreactor into


Table 1Experimental conditions and attained maximum viable cell density and specific antibody productivity (ultrasonic filter fermentations)

bcl-2 cells pEF cellsParameter

Initial cell concentration 7.9×105 cells ml−1 5.7×105 cells ml−1

Culture volume 1 l 1 lAt inoculation At inoculationStart of perfusion

Perfusion rate 0.71–2.03 day−1 0.71–2.03 day−1

Power: 5 W, output 70% Power: 5 W, output 70%Settings for the ADI 1015 Bio Sep (ultrasonicseparator)

On–off cycles: 300 s/3 s On–off cycles: 300 s/3 sFrequency: :2.12 MHz (automatic tuning)Frequency: :2.12 MHz

(automatic tuning)Duration 184 h 188 hNotes Steps of perfusion were performed in the same

way as in the bcl-2 run1.21×107 cells ml−1 0.60×107 cells ml−1Maximum viable cell density

0.81 mg per 106 cells h−10.31 mg per 106 cells h−1Specific antibody productivity

the region below the resonator. The harvestpump, operating at a lower level, draws the cellsuspension into the resonator. High frequencysound waves from the frequency generator enterthe resonance chamber via the solid state trans-ducer and cause the cells to aggregate in theresonance field. The resonance control system pe-riodically turns off the acoustic field to allow theaggregates to settle.

These aggregates are flushed back into thebioreactor by the continuously recirculating cellsuspension while clarified medium is withdrawnfrom the top of the resonator by the harvestpump. The frequency of these on–off cycles isadjusted according to cell types and cell densities.Fresh medium can be added to the culture tomaintain a constant culture volume, using a levelsensor and controller to control a medium supplypump.

Using a similar device, Trampler et al. achievedcell retentions of up to 99% in a high-densityperfusion culture of hybridoma cells without mea-surable effects on viability.

This type of reactor itself offers particular ad-vantages compared to systems based on the use offiltration surfaces that may foul or block. Inparticular, the installation is simple and offersnegligible need for either increase in working vol-ume or compromise of the system integrity orhomogeneity.

In this work we have sought to exploit togetherfor the first time the potential advantages of amore robust cell line transfected with the bcl-2gene when cultured in such an intensified perfu-sion process. Also we compare the novel ultra-sonic cell filter as a cell retention device with thewell-established spin filter under similarconstraints.

2. Materials and methods

2.1. Reactor description

A standard LSL Biolafitte (LSL-Group, Luton,UK) hemispherical-bottomed glass bioreactor wasused (inner diameter 105 mm, height 260 mm),with the configuration seen in Fig. 1. Mixing wasprovided by a top driven, 3-bladed marine im-peller with a diameter of 46 mm rotating at 110rpm. Oxygen was supplied bubble-free by im-mersed silicone tubing (inner diameter 3.2 mmwall thickness 0.8 mm), dissolved oxygen alwaysbeing controlled at 50% air saturation. The pHwas controlled at 7.1 by automatic addition ofCO2 or NaOH (0.2 M). During spin filter fermen-tations the silicone tube was wound around thefittings in the bioreactor in a length of 4.0 m.Ultrasonic filter fermentations were carried outusing a special stainless steel holder for the sili-


Table 2Experimental conditions and attained maximum viable cell density and specific antibody productivity (spin filter fermentations)

bcl-2 cells pEF cells

Initial cell con- 2.0×105 cells ml−11.5×105 cells ml−1

centration1.5 lCulture volume 1.5 l

Start of perfu- After 44 h batch timeAfter 67 h batch timesion

Perfusion rate 0.35–2.27 day−1 0.35–2.27 day−1

Duration 304 h 257 hNotes Culture had to be stopped because spin filter Perfusion was started early because the viability had already

was blocked decreased after 44 hDuration of each step of the perfusion rate was the same asin the bcl-2 run

1.58×107 cells ml−1Maximum viable 0.60×107 cells ml−1

cell density0.93 mg per 106 cells h−10.29 mg per 106 cells h−1Specific antibody

productivity

cone tube. It was mounted between the fittings inthe bioreactor and supported a length of 3.0 m oftube. The cylindrical spin filter had a diameter of4 cm and a height of 9 cm. Its pore size was 25mm, and the immersed filter surface area wasapproximately 107 cm2.

2.2. Cell line and culture conditions

The cell line used throughout this work wasTB/C3 which is a mouse-mouse, NS1-derived hy-bridoma which produces immunoglobulin G(IgG) monoclonal antibody (MAb) specific to thehapten Cg2 domain in the Fc region of humanIgG. The cells were transfected either with thebcl-2 gene (TB/C3 bcl-2) or the negative transfec-tion vector, which gives the control cell line (TB/C3 pEF). The TB/C3 bcl-2 cell line overexpressesthe ‘anti-apoptotic’ human bcl-2 protein. Themedium used throughout the experiments wasRPMI 1640 (Gibco, UK) supplemented with 5%(v/v) foetal calf serum (FCS). A volume of culture(from routine passage) containing enough cells toinoculate a bioreactor at a cell number of 2×105

cells ml−1 was taken and resuspended in freshmedium after centrifugation (10% of spentmedium was carried over in the inoculum). Thefollowing additional supplements were used:

� Penicillin–streptomycin 1% (v/v) at 5000 Unitsml−1 (Sigma-Aldrich, UK)

� Meat peptone 2.5 mg ml−1 (Sigma-Aldrich,UK).

Cell concentration was determined by haemocy-tometer count (improved Neubauer–Haemocy-tometer), and the trypan blue exclusion method wasused to distinguish viable from non-viable cells.

2.3. Fluorescence microscopy: acridineorange–propidium iodide staining

This method was used in addition to the trypanblue staining method in order to determine thedistributions of viable, apoptotic, necrotic andsecondary necrotic cells. Each cell suspensionsample was mixed with an equal volume of stain-ing solution containing 10 mg ml−1 acridine or-ange (AO) and 10 mg ml−1 propidium iodide (PI).At higher cell densities, the samples were diluted(e.g. 1:10–1:20 at cell densities ]107 ml−1). Themixture was loaded into an improved Neubauerrhodium-coated haemocytometer and cell num-bers determined under fluorescence microscopy(excitation by UV light). By examination ofcolour and chromatin morphology the cells wereclassified and the results expressed as percentagesof the total cells. Cells which exclude PI appeared


Fig. 2. Perfusion rate, cell growth and specific consumption rates during ultrasonic filter fermentations: (A) bcl-2 cells, (B) pEF cells.

green. If such cells showed a diffuse chromatinthey were counted as viable cells. Otherwise, iftheir chromatin was condensed they were scoredas early apoptotic cells. Red cells included PI andwere counted as late apoptotic if their chromatinwas condensed or as necrotic if they showed adiffuse chromatin (Simpson et al., 1997).

2.4. Other determinations

The concentration of monoclonal antibody wasdetermined using a sandwich type ELISA. Cap-ture antibody: Human IgG (Sigma I4506); pri-mary antibody (standard): Mouse anti-humanIgG (Jackson Affinipure number 209-005-082,


Fig. 3. Perfusion rate, cell growth and specific consumption rates during spin filter fermentations: (A) bcl-2 cells, (B) pEF cells.

from Stratech Scientific); secondary antibody:Sheep anti-mouse IgG, horseradish peroxidaseconjugate (Sigma A 6782).

Glucose in the culture supernatant was deter-mined each day of the experiment using an enzy-matic test kit (Glucose/GOD-Perid® Method,BOEHRINGER MANNHEIM).

For the determination of glutamine in the cul-

ture supernatant a colorimetric assay kit (Sigma,Stock No. GLN-2) was used. The assay is basedon the reductive determination of L-glutamine bya proprietary enzyme. Quantitation is accom-plished by linking a dye directly to the reductivereaction. The reaction is specific for L-glutamineand does not cross-react with other amino acids orammonia.


Fig. 4. Levels of viable, necrotic and apoptotic cells during ultrasonic filter fermentations: (A) bcl-2 cells, (B) pEF cells (determinedby fluorescence microscopy after acridine orange–propidium iodide staining).

2.5. Calculations

Cell separation efficiency:

E=�

1−cH

cF

�×100% (1)

where E is the cell separation efficiency; cH is thecell concentration in harvest; and cF is the cellconcentration in bioreactor.

Specific consumption rates: A specific substrateconsumption rate can be obtained from the mate-rial balance as follows:

d(VS)dt

=FiSi−VrS−FoSo (2)

where V is the culture volume; Si, So are theconcentration of substrate (inlet/outlet stream);and rS is the consumption rate.

The culture volume of the reactor was keptconstant; therefore Eq. (2) can be written as:

dSdt

=D(Si−S0)−rS

D=Fi

V=

Fo

Vperfusion rate (3)


Fig. 5. Levels of viable, necrotic and apoptotic cells during spin filter fermentations: (A) bcl-2 cells, (B) pEF cells (determined byfluorescence microscopy after accordion orange–propidium iodide staining).

Rearranged in terms of the consumption rate thisgives the following:

rs=D(Si−So)−dSdt

(4)

To obtain the specific consumption rate, rS has tobe divided by the cell concentration:

rS,sp=1

ci

�D(Si−So)−

dSdtn

(5)

where rS,sp is the specific consumption rate.

dS:DS=S1−S0

& where ci is the average cell concentration intime interval t1− t0.

dt:Dt= t1− t0

where S1, S0 are the substrate concentrations attimes t0 and t1.

3. Experimental work

Cultures using the ultrasonic filter were startedat a higher cell concentration than the spin filtercultures in order to reduce the batch time needed


Table 3Separation efficiency of the ultrasonic filter

Retention viable (%) Retention necrotic (%)Cell line Retention of total cells (%) Retention apoptotic (%)

99.3–99.699.999.8–99.9 97.8–99.3bcl-2 cells99.5–99.8 99.5–99.9 99.2–99.6 98.7–99.5pEF cells

to reach the maximum cell concentration. Spinfilter fermentations were carried out in order toobtain comparable data from a different retentionsystem. Bcl-2 and pEF cells were run using thesame stepwise increase of perfusion rates in orderto have a direct comparison between the two celllines. The experimental conditions for the fermen-tations using the ultrasonic filter are summarisedin Table 1 and for the fermentations using thespin filter in Table 2.

4. Results and discussion

4.1. Cell growth

Both retention systems show similar resultsconcerning the maximum cell density. Using thesame perfusion rate strategy in both systems, thebcl-2 transfected cells reach double the viable celldensity of the control cells (pEF) as depicted inFig. 2 (ultrasonic filter) and Fig. 3 (spin filter).

As can be seen in Figs. 4 and 5, there is noevident difference in the levels of viable, necroticand apoptotic cells between the ultrasonic and thespin filter runs. The time scales differ only by thetime required for the cell concentration to risefrom the inoculum concentration used in the spinfilter (1.5–2.0×105 ml−1) to that used with theultrasonic filter (5.7–7.9×105 ml−1). This indi-cates that ultrasonic treatment at the power levelsapplied has no negative effect on the viability ofthe tested cell lines used. However, a clear differ-ence can be seen between the two cell lines in themaintenance of viability. In both cases (ultrasonicand spin filter), bcl-2 cells maintain a viabilityover 90% and apoptotic cells are less than 5%,whereas with the pEF control cells the viabilitygoes down almost to 70% and apoptosis reacheslevels between 10 and 20%. Death rates are clearlyreduced by the bcl-2 transfection.

Comparing pEF and bcl-2 in batch culture,Simpson et al. (1997), found that bcl-2 cells re-tained higher levels of membrane-intact (MI) cells(a measure of viability, by trypan blue exclusion)than pEF cells. In this pEF culture, MI cellsrapidly declined just after reaching the maximumcell density whereas the bcl-2 cells entered a sta-tionary phase after which the rate of loss of MIcells was slower.

Figs. 2 and 3 also show that the specific glucoseconsumption rate of the pEF cells is double thatof the bcl-2 cells. The glucose concentrations ofthe pEF cultures have at no time been lower thanin the bcl-2 cultures (data not shown) but appar-ently the growth of the pEF cells is limited byglucose. However, this deprivation does not affectthe growth of the bcl-2 cells.

4.2. Cell retention of the ultrasonic filter

The ultrasonic filter bioreactor was operated atflow rates between 0.7 and 2.0 l day−1. Separa-tion efficiencies were consistently high, varyingfrom 99.5 to 99.9% (Table 3) and showing nodiscernible dependency on the applied flow rates.

Trampler et al. (1994), using a 32 ml resonatorin a high density hybridoma culture, reached sep-aration efficiencies ranging from 90 to 96% atflow rates from 0.35–1.0 l h−1. At flow rateshigher than 1.0 l h−1 the separation efficiencydistinctly decreased, dependant on the flow rate.

Table 3 also shows the different separationefficiencies for viable, necrotic and apoptotic cells.One can see that viable cells are better retainedthan both necrotic and apoptotic cells. However,the difference in the retention of each fraction isquite small, ranging only from 98 to 99.9%.Trampler et al. (1994) found that viable cells wereselectively retained at up to 3% higher efficiencythan nonviable cells. It is assumed that such a


Fig. 6. Formation of clumps during ultrasonic filter fermenta-tion, pEF cells, cultivation time 140 h: (A) Clumps of non-vi-able cells (trypan blue staining, magnification: ×100). (B)Clumps comprising viable and non-viable cells. The structure

selective retention results from size and compress-ibility differences between viable and nonviablecells, both of which influence the acoustic forceson a particle.

It would be desirable to be able to preventaccumulation of dead cells in the bioreactor byretaining viable cells preferentially. However, con-sidering the relatively low total cell number in theharvest, the selective removal of dead cells clearlyis a negligible effect.

During the pEF fermentation using the ultra-sonic filter, clearly visible clumps were formingafter cell concentrations reached 106 cells per ml.Size measurements showed clump diameters up to1 mm. Microscopic examination showed an amor-phous mass in which no distinct cells were visible.Trypan blue stained the clumps dark blue. Somesmaller clumps are shown in Fig. 6: Panel Bdepicts single cells which can be clearly distin-guished from each other, whereas in Panel C ofFig. 6 cells are already loosing their structure.Possibly, the cells are forming aggregates insidethe resonator and, after returning to the vessel,they do not always break up completely. If largerclumps enclose viable cells, it is quite likely thatthey are not supplied sufficiently with nutrientsand oxygen.

The question as to whether cells are dyingbecause of the formation of aggregates or whetherthe larger clumps are formed only from deadmaterial, e.g. cell debris and protein, can not beanswered by these examinations and needs furtherinvestigation.

4.3. Production of monoclonal antibody

Both bcl-2 and pEF cells show similar produc-tivities in each reaction system (see Tables 1 and2). Within the limits of accuracy of the ELISAassay there is no indication that productivity is

of the individual cells is still visible (trypan blue staining,magnification: ×400). (C) Clumps showing mainly non-viablecells. Cell structures seem to disintegrate into an amorphousmass (trypan blue staining, magnification: ×400).


Fig. 7. Accumulated amount of antibody during the fermentations.

affected by ultrasonic treatment. Fig. 7 depicts theaccumulated antibody in each run. Tables 1 and 2both show a 3-fold higher productivity of pEFcells compared to that of bcl-2 cells. Simpson etal. (1997) found that the Mab titres of bcl-2 batchcultures were 38% higher than in similar pEFcultures. They observed that bcl-2 cells producedat a rather consistent and higher rate than pEF.However, in their experiments, the productivity ofpEF cells increased in association with the onsetof cell death, even exceeding the productivity ofbcl-2 cells during phases of high death rates. Thisnegative association between productivity andgrowth is therefore probably due in part to thepassive release of Mab stored in cytoplasmic vesi-cles as cells die and lose their membrane integrity.It is doubtful if such product is complete inrespect of post-translational modifications. Thus,the higher productivity of the pEF cells is proba-bly due to the higher death rate of the pEF cells.Other possibilities are that repeated subculturingof the bcl-2 cell line has favoured the selection ofa clone which is on the one hand highly resistantto culture stress but on the other hand has a lowerproductivity than the original cell line, or that

over-expression of bcl-2 compromises the productprotein synthesis of the cell, perhaps through aribosomal limitation.

5. Conclusions

� The ultrasonic filter is an effective tool forachieving high retention levels of cells in perfu-sion culture

� At the recommended applied power levels,there is no indication that ultrasonic treatmentaffects the viability or productivity of the celllines tested.

� A selective retention of viable cells is detectablebut its effect on the removal of non-viable cellsis negligible.

� The involvement of ultrasonic treatment in theformation of clumps needs furtherinvestigation.

� The transfection of cells with the anti-apoptosisgene bcl-2 clearly reduces the susceptibility toculture stress. Cell lines based on this approachmay show significant advantages in intensivesystems using cell retention.


References

Al-Rubeai, M., Singh, R.P., Goldman, M.H., Emery, A.N.,1995. Death mechanisms of animal cells in conditions ofintensive agitation. Biotechnol. Bioeng. 45, 463–472.

Griffith, B., 1990. Perfusion systems for cell culture technol-ogy. In: Lubiniecki, A.S. (Ed.), Large-Scale MammalianCell Culture Technology. Marcel Dekker, New York, pp.217–250.

Itoh, Y., Ueda, H., Suzuki, E., 1995. Overexpression of bcl-2,apoptosis suppressing gene: prolonged viable culture pe-riod of hybridoma and enhanced antibody production.Biotechnol. Bioeng. 48, 118–122.

Kerr, J.F.R., Wyllie, A.H., Currie, A.R., 1972. Apoptosis: abasic biological phenomenon with wide ranging implica-tions in tissue kinetics. Br. J. Cancer. 26, 239–257.

Phylis, W.S., Trampler, F., Sonderhoff, S., Groeschl, M.,Kilburn, D.G., Piret, J.M., 1995. Batch and semicontinu-ous aggregation and sedimentation of hybridoma cells byacoustic resonance fields. Biotechnol. Prog. 11, 146–152.

Simpson, N.H., Milner, A.E., Al-Rubeai, M., 1997. Prevention

of hybridoma cell death by bcl-2 during sub optimalculture conditions. Biotechnol. Bioeng. 54, 1–16.

Singh, R.P., Al-Rubeai, M., Gregory, C.D., Emery, A.N.,1994. Cell death in bioreactors: a role for apoptosis. Bio-technol. Bioeng. 44, 720–726.

Singh, R.P., Emery, A.N., Al-Rubeai, M., 1995. Enhancementof survivability of mammalian cells by overexpression ofthe apoptosis-suppressor gene bcl-2. Biotechnol. Bioeng.52, 166–175.

Singh, R.P., Finka, G., Emery, A.N., Al-Rubeai, M., 1996a.Apoptosis and its control in cell culture systems. Cytotech-nology 23, 87–93.

Singh, R.P., Al-Rubeai, M., Emery, A.N., 1996b. Apoptosis:exploiting novel pathways to the improvement of cellculture processes. Genet. Eng. Biotechnol. 16, 227–251.

Trampler, F., Sonderhoff, S., Phylis, W.S., Kilburn, D.G.,Piret, J.M., 1994. Acoustic cell filter for high cell densityperfusion culture of hybridoma cells. Bio/Technol. 12,281–284.

Tsujimoto, Y., Cossmann, J., Jaff, E., Cruce, C.M., 1985.Involvement of the bcl-2 gene inhuman follicularlymphoma. Science 228, 1440–1443.

.

Documents

A comparison of intensive cell culture bioreactors ... · mechanism of cell death during the culture of many industrially important mammalian cell lines. The levels of apoptosis in