Vlak J.M., de Gooijer C.D., Tramper J., Miltenburger H.G. (Eds.) Insect Cell Cultures: Fundamental and Applied Aspects

Подождите немного. Документ загружается.

182

oxygen transfer rates with low and homogeneously dis-

tributed shear compared to conventional stirred reac-

tors or simple bubble columns. Aspect ratio and super-

ficial gas velocity influence reactor performance and

they must be assessed at each higher scale using as a

criterion the meeting of desired oxygen transfer rates.

The ratio of downcomer to riser cross sectional areas is

critical to mixing and oxygen transfer in an airlift reac-

tor, and as this ratio is increased the volumetric mass

transfer coefficient and the superficial velocity

of the air decrease. The successful cultivation of Sf9

cells in a 21-liter airlift reactor for the production of

macrophage colony stimulating factor was reported by

Maiorella et al. (1988). Using a new serum-free medi-

um formulation containing Pluronic F-68, plus a close

control of the air bubble diameter (0.5-1.0 cm) to avoid

coalescence, they optimized the process to a maximum

cell density of and a protein titer of

performances comparable to those obtained

initially in 100-ml spinners. In a 14-liter airlift reactor

operating at an aeration of 0.06 vvm with 1-cm air bub-

bles, a kinetic study of Sf9 growth and recombinant

galactosidase production showed that a very high cell

density of was possible and 0.33 mg

protein per ml of serum-free medium (SF-900) was

reached, i.e., about 37% of the level attained in shake

flasks (King et al

1992). A bigger scale of cultiva-

tion proved possible when Weiss and colleagues (1988)

used a microbial 40-liter concentric tube airlift fermen-

tor to arrive at a cell density of

from an inocolum level of over

11 days, while Schlaeger & co-workers (1994) adapt-

ed Tn5 cells to suspension growth in 60-liter airlifts in

serum-free medium containing Pluronic F-68.

Conventional stirred tank systems have been used

for large-scale insect cell culture ranging from volumes

of 3 liters to over 150 liters (Weiss et al

1989; Jain

et al

1991; Barkhem et al

1992; Guillame et al

1992; Bédart et al

1994b). The basic attraction of a

stirred tank reactor system is its wide availability and

suitability for multipurpose use (e.g., in a pilot plant

where R&D tests may alternate among various micro-

bial and animal cell systems for production of biolog-

icals) and the long-standing experience in design and

understanding of scale-up principles in microbial fer-

mentation. The fear of hydrodynamic shear-induced

damage to insect cells from air bubbles has motivated

the development of bubble-free aeration techniques in

stirred tanks which can be fitted with semipermeable

membrane (silicon) tubing for oxygenation. Exam-

ples include the successful cultivation of Mamestra

brassicae cells (Miltenburger & David, 1980) and Sf9

cells (Eberhard & Schügerl, 1987) in a 10-liter reactor

encompassing 15 m of silicon tubing. Further improve-

ments in oxygen transfer can be realized in this system

when the aeration tubing is motorized with a recipro-

cating movement (Graf & Schügerl, 1991). Despite

the demonstration of this system’s feasibility at an

intermediate scale, further scale-up will be limited by

the progressively larger percentage of the reactor vol-

ume that will be occupied by the tubing, its high cost

and the problems of membrane fouling anticipated.

On the other hand, there is encouraging evidence that

large-scale production of insect cells and their recom-

binant protein or viral products can be achieved with

standard agitated reactors after relatively minor mod-

ification (design and configuration of impellers and

aeration orifices, complementation of efficient surface

aeration and non-aggressive sparging in combination

with shear-protective media additives). For example,

a modified 8-liter spinner reactor agitated with mag-

netic bars at 90 rpm and sparged directly with air at

through a conventional ring sparger gave

satisfactory cell densities

in serum-free medium (Weiss et al

1989). The same

workers demonstrated the potential of a 3-liter spin-

filter reactor to produce of recombi-

nant human (Weiss et al

1989). In this

design, the cells are physically segregated from the

sparged medium through a rotating cage, continuous

perfusion of the medium and removal of metabolic

waste by-products is occurring and high cell densities

of the order of are attainable. In a vari-

ation of this technique, removal of spent medium is

continuously effected through a microporous filtration

module or a hollow fiber filtration cartridge, resulting

in production of more than 11 viral polyhedra per cell,

compared to only 9 without perfusion (Klöppinger et

al., 1990). The feasibility of high cell density scale-up

was also demonstrated in a modified CelliGen (New

Brunswick Scientific, Edison, NJ) bioreactor adapted

for perfusion culture (Caron et al

1994). Sparging

was done through a geotextile membrane on the base

of the CelliGen’s characteristic hollow impeller shaft;

0.5-mm bubbles were thus harmless for the cells that

were not exposed to them. When perfused at 1 - 1.5

volume of medium per day, Sf9 cells grew from

to over 3-4 days. The continuous

perfusion was possible through prototype tangential

filtration devices. A more complete discussion ot the

potential of high cell density perfusion cultures is giv-

en elsewhere in this volume (Jäger, 1996). A compar-

183

ison of the performance of a prototype 11-liter stirred

tank bioreactor equipped with a helical ribbon impeller

(Kamen et al

1991) and a modified 4-liter stirred tank

CelliGen bioreactor with two marine impellers and a

special interfacial impeller (to improve surface aeration

from the head space) used by the same group (Caron

et al

1990) showed that the final Sf9 cell densities

attained were comparable, in the vicinity of

Differences in the yield of recombinant

protein production (bovine rotavirus nucleocapsid pro-

tein VP6) were attributed to the use of different media

between the two systems.

The lytic insect cell-baculovirus expression system

imposes limitations on the useful length of a batch

reactor cycle, hence continuous insect cell cultivation

schemes have been proposed in order to reach high and

uniform productivity. Cascades of multiple reactors

have been reported (Kompier et al

1988; Klöppinger

et al

1990; 1991; van Lier et al

1990; 1992; Zhang

et al

1993). In these (semi-) continuous setups two

or three reactors are run in series, the first for growth

of uninfected cells followed by one or two vessels for

cell infection and virus or protein production. The sys-

tems proposed by Klöppinger et al. (1990; 1991) for

Sf9 cells and by Zhang et al. (1993) for Bm5 cells

are in essence semi-continuous (repeated batch culti-

vation of cells and perfusion of medium in the infec-

tion reactor which can be infected several consecutive

times) and achieve adequate and reproducible prod-

uct titers. In contrast, the systems proposed by the

Wageningen group (Kompier et al

1988; van Lier et

1990; 1992) are completely continuous. The effec-

tive production period for polyhedra or

in this scheme sems to be of the order of about 30 days,

beyond which point there is an unrecoverable decline

in productivity attributed to the “passage effect”. This

effect is considered to be due to the progressive enrich-

ment of the serially passaged virus in defective inter-

fering particles (DIP), i.e., deletion mutants interfering

with the multiplication of fully functional polyhedro-

sis virus (Kool et al

1991; Wickham et al

1991).

A concern to use low-passage viral inocula in continu-

ous or semi-continuous bioprocessing schemes and the

greater flexibility of fed-batch cultivation may improve

the outlook of these modes of bioreactor operation.

Monitoring of insect cell growth and baculovirus

infection phases

As in every bioprocessing scheme involving

production-scale microbial or mammalian cell sys-

tems, the ultimate consistency of bioreactor-based

insect cell cultivation can only be assured with ade-

quate means of process monitoring and control in

place. In addition to the traditional bioreactor culti-

vation parameters that are measurable on line on most

vessels, such as temperature, pH and dissolved oxy-

gen (DO), continual (ideally on-line) measurements

of cell density and viability, of concentrations of sub-

strates and byproducts, etc. are important for track-

ing the progress of the cultivation with time and for

establishing simple empirical or model-based control

actions in order to maintain optimal conditions or to

steer the process in the direction most compatible with

process goals. The choice of parameter to be moni-

tored and controlled will be affected by the degree to

which it influences, directly or indirectly, the efficien-

cy of protein production in the insect cell-BEV system

(see above). As with mammalian cell reactors, instru-

mentation of insect cell reactors can encompass both

well-established technological tools for on- or off-line

estimation of respiration rates and quotients (Kamen

& Tom, 1994; Wong et al

1994; Hensler & Agathos,

1994) as well as less conventional accessories, such as

optical density probes (Bédart et al

1994a) or cou-

pling to biochemically relevant analyses of cell popu-

lations, such as flow cytometry (Al Rubeai et al

1994)

or fluorescent probes (Hensler & Agathos, 1994).

The determination of viable and total cell count

is among the most important measurements during

both the growth and the production phase in insect

cell reactors. Cell numbers are typically obtained off-

line by counting trypan blue-stained and unstained por-

tions of cell suspension aliquots in a hemocytometer

or by using a Coulter™ particle counter. Although

still widely practiced in laboratory and industrial sit-

uations, manual counting is labor-intensive and does

not lend itself easily to automation, hence direct phys-

ical approaches are considerably more attractive for

process data logging and control purposes. Measure-

ments of optical density (OD) obtained either off-

line on a spectrophotometer or on-line with an opti-

cal density sensor have been recently reported to

be reliable means of assessing cell concentration in

bioreactor-based insect cell cultures (Bédart et al

1994a). After establishing the validity of a simple

relationship between OD and Sf9 insect cell density,

Bédart et al. (1994a) recognized that the continuous

monitoring of OD would be both useful and feasible

in scale-up, since a new generation of steam steriliz-

able optical sensors have been introduced lately for

on-line OD logging in animal bioreactors (Konstanti-

184

nov et al

1992; Zhou & Hu, 1994). The on-line OD

measurement of cultured Sf9 insect cells was carried

out with a MAX™ Cell Mass Sensor (Cerex Corpo-

ration, Ijamsville, MD) consisting of an optical probe

and a data acquisition system connected to a personal

computer. The fiber optic probe (5 mm light path) was

equipped with a laser diode light source with a spectral

range of 820-850 nm. The system also contained an

integrated debubbler to minimize interference from air

bubbles in the 20-liter bioreactor culture. The useful-

ness of OD monitoring is currently limited to the cell

growth phase only, because it has been found that OD

measurements in baculovirus-infected cultures grossly

overestimate the actual cell density due to changes in

cell size and to the presence of cell debris during the

progress of infection (Bédart et al

1994a). However,

with the high accuracies of on-line OD measurements

for Sf9 cell densities up to reported by

these workers, it is easy to obtain satisfactory estimates

of derivative (calculated) culture parameters, such as

specific growth rates and specific nutrient consump-

tion rates, which could be useful in their own right for

process control.

Indirect measurement procedures linking cellular

growth with aerobic respiration are also promising in

insect cell bioreactors: for instance, Kamen & Tom

(1994) have found that there is a consistent linear rela-

tionship between viable cell density and Produc-

tion Rate (CPR) during Sf9 cell growth over a wide

range of cell concentration values

Accurate on-line estimates of CPR and OUR are easy

to obtain from instrumented bioreactors equipped with

gateway offgas sensors such as a mass spectrometer

(Kamen & Tom, 1994). Offgas analysis can therefore

provide continuous accurate estimates of cell density

during the insect cell growth phase. In addition, OUR

in itself appears to be a useful metabolic parameter for

tracking both growth and infection/protein production

phases in insect cell bioreactors (Wong et al

1994;

Hensler & Agathos, 1994). For exponentially growing

uninfected Sf9 cells growing in a 1-liter stirred tank

reactor it was found that OUR was proportional to cell

density over a range of

(Wong et al

1994). The respiratory quotient (RQ),

defined as the ratio of CPR/OUR, has only recently

been recognized as a potentially valuable indicator of

the physiological state in animal cell culture (Bonarius

et al

1995) and therefore it would be interesting to

assess its utility in monitoring insect cell bioreactor

cultivations (Kamen & Tom, 1994).

In contrast to the growth phase, the monitoring of

baculovirus-infected cells is considerably more chal-

lenging, even though this phase is crucial for tracking

and controlling protein or virus production. Conven-

tionally, viable cell density in the phase of infection

is followed in discrete samples by microscopic obser-

vation in combination with trypan blue dye exclusion,

as mentioned above. In recent work we have evaluated

the progress of infection of Sf9 cells in DO-controlled

250-ml stirred tanks for recombinant

production by means of OUR and epifluorescence

microscopy in addition to trypan blue staining (Hensler

and Agathos, 1994). A combination live/dead fluores-

cent stain system (Molecular Probes, Eugene, OR)

consisted of calcein AM and of ethidium homodimer

stains. The calcein AM stain gave an assessment of

cell viability by virtue of its hydrolysis into a highly

fluorescing (green) product through the action of intra-

cellular esterases, while ethidium homodimer provided

a powerful visual evaluation of the status of infected

cells since this stain penetrates compromised cellular

membranes and intercalates with the cell’s DNA pro-

ducing a bright orange fluorescent glow. This combina-

tion live/dead staining procedure gave a better assess-

ment of the gradual loss of viability of infected Sf9

cells compared to trypan blue exclusion, which tended

to overestimate viability in the early stages of infec-

tion and to underestimate it 3-4 days post-infection.

While this assessment was done by discrete sampling

off-line, the samples required no special processing

(fixing etc), hence fluorometric measurements can be

potentially adapted to continuous monitoring on-line.

OUR has also been found useful in following the

progress of the infection process given its transiently

increasing profile (Hensler & Agathos, 1994) as well

as the recently noted inverse correlation between the

rate of rise in the specific OUR and the cell density at

the time of infection (Wong et al., 1994). Based on this

observation Wong and co-workers suggest that OUR

measurements can be used to monitor the efficiency of

a batch infection process. Such a series of OUR mea-

surements during the early stages of the infection may

indicate whether the infection will proceed as predict-

ed by comparing the culture OUR with the expected

rate of specific OUR rise, and corrective interventions

such as medium replenishment or nutrient feeding can

then be imposed in order to restore the productivity, if

the rate of increase in specific OUR is found to reflect a

potentially low-yielding infection (Wong et al

1994).

Yet another useful means of monitoring baculovirus

infection of cultured insect cells is flow cytometric

185

analysis of unprocessed culture samples (Al Rubeai

et al

1994). In this work the orthogonal scattering

properties were found to be linearly related to the per-

centage of infected Sf9 cells as assessed by optical

microscopy. In addition, flow cytometric monitoring

of cellular DNA identified the infectivity of the cells

and the subsequent development of viral particles much

earlier than visually possible.

The evaluation of optimal harvesting time is also

a matter of careful consideration. It must be aided by

reliable means of product monitoring and may well

depend on the technique used. Hence methods of ade-

quate continuous or discrete measurement of product

formation in bioreactors are highly desirable. In our

own work, gel electrophoretic, chromogenic (ONPG)

or fluorometric methodologies for a model recombi-

nant protein, bacterial gave varying

indications of peak gene expression for the same Sf9

culture (Hensler and Agathos, 1994). SDS-PAGE gel

profiles indicated a peak in expression at 48-72 hours

post infection, whereas the ONPG assay together with

a fluorometric assay of unprocessed cells stained with

fluorogenic substrates such as

(ImaGene Green, Molecular Probes, Eugene, OR)

indicated a peak later in the infection (96 hours). The

power of visualization of cellular con-

tent through the same fluorometric methodology was

also reported by others (Al Rubeai et al., 1994).

Bioreactor mode of operation in relation to

nutritional requirements: medium replenishment or

nutrient addition strategies

We noted above that the desirability of maximal pro-

ductivities by operating high cell density cultivations

of insect cells must be tempered with the realization

that nutritional control should be exercised in the infec-

tion phase, in order to overcome the apparent conflict

between high-density infection and good recombinant

protein formation.

During insect cell infection, cell division stops but

other cell physiological activities continue, together

with transcription and translation that is now com-

mandeered towards viral multiplication and viral gene

expression. Hence, for high virus or protein production

to be accomplished, it is vital that the infected cell’s

machinery be kept under optimal physiological con-

ditions, free of oxygen or nutritional limitations and

of potential inhibitions from accumulating metabolic

products. This realization has been supported by recent

studies showing that by resuspending cells in fresh

medium upon infection at high density and by main-

taining aeration, specific productivity can be restored

to that of low-density infection or even enhanced (Lind-

say & Betenbaugh, 1992;Reuveny et al., 1993a;Zhang

et al

1993; 1994b; Hensler & Agathos, 1994). For

example, an increase of Sf9 density at infection from

to led

to a 17 fold reduction in specific productivity of

galactosidase, yet resuspension in fresh medium after

high-density infection restored specific productivity

to 100% compared with the low-density culture and

allowed a 26-fold increase in protein yield compared

to infection in spent, unaerated medium (Lindsay &

Betenbeugh, 1992). Similarly, by renewing the serum-

free medium before infection at Sf9 cell densities of

2 and a similar

specific productivity was obtained to that observed at

in spent medium, whereas infec-

tion at without complete medium

renewal had previously led to a 3-fold decrease in pro-

tein per cell compared to that observed at

(Hensler & Agathos, 1994). In this case

nutrient limitation was ascertained as the root cause

of the unsatisfactory performance at high cell densi-

ty infection and the medium renewal stategy led to

a doubling of maximum volumetric productivity. In

addition, measurements of nutrient specific consump-

tion rates showed that the utilization of glucose and

glutamine increased in infected Sf9 cells by 70% and

64% respectively, compared to the rates in exponen-

tially growing uninfected cells (Hensler & Agathos,

1994). In a series of detailed nutritional studies, the

limitations of recombinant protein (epoxide hydrolase)

production in Sf9 cells by oxygen, glucose and glu-

tamine were demonstrated and supplementation with

these nutrients, together with DO control at 35% satu-

ration, increased significantly the protein specific pro-

ductivity at the high cell density of

(Wang et al., 1993a, 1993b). Although no amino acid

was found to be protein formation rate-limiting dur-

ing the infection phase at the lower cell densities of

(Wang et al

1993b), there

are indications that some amino acids may indeed be

rate-limiting at high cell density infection, given the

increases noted in their specific consumption rates after

infection in comparison to their pre-infection values

(Wong et al

1994). Along the same lines, direct intro-

duction of fresh medium after viral infection increased

productivity of recombinant HIV antigen CD4 at com-

parable cell densities (Lazarte et al

1992). In a study

of recombinant protein or glucocere-

186

brosidase) formation by Sf9 cells under oxygen suffi-

ciency, protein production was tripled when the spent

medium was replaced with fresh one before infection

(Reuveny et al., 1993a). In the same work, when the

fresh medium was supplemented with either glucose

alone or glucose, glutamine and yeastolate a 25% or a

100% increase in recombinant protein expression was

observed, respectively.

Occasional reports of no direct effect of cell density

levels nor of nutrient replacement strategies on protein

productivity (Kioukia et al., 1995; Zhang et al., 1994b)

may indicate that other simultaneous influences should

also be examined for their role on the cells’ infectivity

and production potential at key points in the popula-

tion’s development. Notably, the difference in infec-

tivity in different phases of the insect cells’ growth

curve may be due to the difference in cell cycle distri-

bution between the G1, S and and G2/M phases and

the higher susceptibility of S-phase cells to baculovirus

replication (Kioukia et al

1995). Also if nutrient lim-

itation is absent at the end of exponential growth, there

is no advantage of medium replenishment for protein

productivity (Zhang et al., 1994b).

In sum, the decline of recombinant protein yield

observed in cultures at late growth stage (high cell

density) can be reversed partially or totally by one

of the following stategic interventions: (a) by com-

pletely replacing spent medium with fresh before or

after infection (Caron et al

1990; Lindsay & Beten-

baugh, 1992; Hensler & Agathos, 1994), (b) by adding

fresh medium directly to the infected sulture (Lazarte et

al., 1992) or (c) by simultaneous medium replacement

and nutrient supplementation (Reuveny et al

1993a).

Along similar lines, the promise of perfusion culture

of insect cells for recombinant protein production at

high cell densities (Caron et al

1994; Deutschmann &

Jäger, 1994) is largely based on the continuous replace-

ment op spent medium with fresh.

Feeding stategies designed for consistently high

protein production should ideally be as simple as pos-

sible, especially if large volumes of mediuum are

involved. The infection and post-infection stage of

high density Sf9 cell culture has been studied recently

with respect to optimal production of a heterologous

protein by focusing on the dynamicss of the protein

as it is being expressed. In a study of the produc-

tion of a recombinant soluble immunoglobulin recep-

tor (IgEr), the concentration of the protein increased

for 4-5 days post-infection but it declined after day 6

(Kirn & Familetti, 1994). This group demonstrated that

IgEr degradation could be minimized by addition of

protein hydrolyzates early after baculovirus infection:

by supplementing the serum-free ExCell 401 medium

with 5 mg/ml meat hydrolyzate or casein hydrolyzate

at 24 hours post infection, a 4-fold increase in IgEr

was achieved over unsupplemented cultures or cul-

tures supplemented prior to infection (Kirn & Familet-

ti, 1994).

In a further recent development, Bédart & co-

workers (1994b) have demonstrated that specific pro-

tein yields in normally low-producing late exponential

phase cultures of Sf9 cells can be restored to high levels

by the one-time addition of an ultrafiltrate of yeasto-

late and a concentrated mixture of amino acids; this

can lead to high volumetric yields of

galactosidase in the serum-free medium SF900II (Gib-

co/Life Technologies, Grand Island, NY), which sup-

ports among the highest Sf9 cell densities in batch

culture The findings of the last

two studies show that exhaustion of one or more nutri-

ents rather than accumulation of inhibitory byproducts

may be the main limiting factor in recombinant pro-

tein production by Sf9 cells in batch culture. From a

process development point of view, this strategy sim-

plifies large-scale protein production because it allows

operations of growth and infection of high cell density

without requiring a major step of cell-liquid separation

for medium replacement.

Concluding Remarks

Despite the variety of bioreactor designs that have been

studied and used in acedemic work and commercial

R&D, no single system has proven so far to be uni-

versally superior in productivity and scalability. These

are attractive features in most attached and suspen-

sion growth systems even though there is a tendency

to favor the latter given their simplicity and the long-

standing experience associated with them from micro-

bial fermentation. The recent insights concerning the

culturability of insect cells even under conditions of

direct aeration in stirred-tank and airlift reactors and

the rationalization of the influences of oxygen, other

nutrients, cell density, multiplicity of infection, etc. on

productivity, should help demystify insect cell culture

in bioreactors. It is especially encouraging that micro-

bial pilot plant reactors of 75 and 280 liters can be

retrofitted in terms of temperature and agitation con-

trol, impeller type (Rushton turbines of hydrofoils),

versatile aeration capability through both head-space

and open-pipe sparging, for successful cultivation of

187

insect and mammalian cells in a rapid and inexpen-

sive manner (Junker et al

1994). Thus it may well

be that large-scale insect cell culture technology will

be progressively practiced through the use of reactors

flexibly designed for both microbial and animal cell

cultivation. Nonetheless, reactor designs for attached

or immobilized insect cell growth (including cell reten-

tion devices for perfusion culture) may also be appeal-

ing alternatives for processes where the scale-up of

production capacity may revolve around very strin-

gent requirements of the particular cell line (attachment

dependence, very high densities on or in particular sup-

port matrices, protection against physical stresses, ease

of medium exchange, etc). The future of bioprocess-

ing with insect cells can be anything but bleak given

the present wealth of options in both equipment and

operation stategy.

Acknowledgments

I wish to thank Helene Skikos Agathos for valuable

assistance in completing the typed manuscript.

References

Agathos SN(1993) The future of cell culture engineering. In: Goosen

MFA, Daugulis AJ & Faulkner P (eds) Insect Cell Culture Engi-

neering, pp. 221–240. Marcel Dekker, New York.

Agathos SN (1994) Large scale insect cell production. In:

Maramorosch K & McIntosh AH (eds) Insect Cell Biotechnology,

pp. 89–103. CRC Press, Roca Raton.

Agathos SN, Jeong Y-H & Venkat K (1990) Growth kinetics of free

and immobilized insect cell cultures. Ann. N.Y. Acad. Sci. 589:

372–398.

Al Rubeai M, Kioukia N & Emery AN (1994) Monitoring of insect-

baculovirus culture processes by flow cytometry. Presented at Cell

Culture Engineering IV, Engineering Foundation Conferences,

San Diego, CA 7-12 March 1994 (Abstract WE3, Cell Culture

Engineering IV Program and Abstracts, p. 86).

Archambault J, Robert J & Tom RL (1994) Culture of immobilized

insect cells. Bioprocess Eng. 11: 189–197.

Barkhem T, Carlsson B, Danielsson A, Norinder U, Frieberg H &

Öhman L (1992) Production in a 100 liter stirred tank reactor of

functional full length human thyroid receptor b1 in Sf-9 insect

cells using a recombinant baculovirus. In: Vlak JM, Schlaeger

E-J & Bernard AR (eds) Baculovirus and Recombinant Protein

Production Processes, pp. 235–246. Editiones Roche, Basel.

Bédart C, Jolicoeur M, Jardin B, Tom R, Serret S & Kamen A (1994a)

Insect cell density in bioreactor cultures can be estimated from

on-line measurements of optical density. Biotechnol. Techniques

8: 605–610.

Bédart C, Kamen A, Tom R & Massie B (1994b) Maximization of

recombinant protein yield in the insect cell/baculovirus system

by one-time addition of nutrients to high-density batch cultures.

Cytotechnology 15: 129–138.

Bonarius HPJ, de Gooijer CD, Tramper J & Schmid G (1995) Deter-

mination of the respiration quotient in mammalian cell culture in

bicarbonate buffered media. Biotechnol. Bioeng. 45: 524–535.

Caron AW, Archambault J & Massie B (1990) High-level recom-

binant protein production in bioreactors using the baculovirus-

insect cell expression system. Biotechnol. Bioeng. 36: 1133–

1140.

Caron AW, Tom RL, Kamen AA & Massie B (1994) Baculovirus

expression system scaleup by perfusion of high-density Sf-9 cell

cultures. Biotechnol. Bioeng. 43: 881–891.

Chalmers JJ (1994) Cells and bubbles in sparged bioreactors.

Cytotechnology 15: 311–320.

Chung IS, Taticek RA & Shuler ML (1993) Production of human

alkaline phosphatase, a secreted glycosylated protein, from a

baculovirus expression system and the attachment-dependent cell

line Trichoplusia ni BTI-Tn 5B1-4 using a split-flow, airlift biore-

actor. Biotechnol. Progress 9: 675–678.

de Gooijer CD, van Lier FLJ, van den End EJ, Vlak JM & Tramper

J (1989) A model for baculovirus production with continuous

insect cell cultures. Appl. Microbiol. Biotechnol. 30: 497–501.

Depinto W & Familetti PC (1994) Increased expression of recombi-

nant proteins from baculovirus infected cultures using Tn-5B1-4

(High Five) insect cells. In Vitro 30: 722 (Presented at the 1994

Cell and Tissue Culture Congress, Research Triangle Park, NC,

4-7 June 1994).

Deutschmann SM & Jäger V (1994) Optimization of the growth

conditions of Sf-21 insect cells for high-density perfusion culture

in stirred-tank bioreactors. Enzyme Microb. Technol. 16: 1133–

1140.

Eberhard U & Schügerl K (1987) Investigations of reactors for insect

cell culture. Develop. Biol. Standard. 66: 325–330.

Fleischaker RJ & Sinskey AJ (1981) Oxygen demand and supply in

cell culture. Eur. J. Appl. Microbiol. Biotechnol. 12: 193–197.

Goldblum S, Bae Y-K, Hink WF & Chalmers J (1990) Protective

effect of methylcellulose and other polymers on insect cells sub-

jected to laminar shear stress. Biotechnol. Progress 6: 383–390.

Graf H & Schügerl K (1991) Influence of the reciprocating move-

ment on the performance of a membrane aeration system in insect

cell cultures. Biotechnol. Techniques 5: 91–94.

Guillame JM, Soria HM, Couteault N, Hurwitz DR, Multon MC &

Crespo A (1992) Insect cell fermentation scale-up for recombi-

nant protein production. In: Vlak JM, Schlaeger E-J & Bernard

AR (eds) Baculovirus and Recombinant Protein Production

Processes, pp. 255–261. Editiones Roche, Basel.

Hensler WT & Agathos SN (1994) Evaluation of monitoring

approaches and effects of culture conditions on recombinant pro-

tein production in baculovirus-infected insect cells. Cytotechnol-

ogy 15: 177–186.

Hensler W, Singh V & Agathos SN (1994) Sf9 insect cell growth

and production in serum and serum-free media.

Ann. N.Y. Acad. Sci. 745: 149–166.

Hink WF (1982) Production

Autographa californica nuclear poly-

hedrosis virus in cells from large-scale suspension cultures. In:

Kurstak E (ed.) Microbial and Viral Pesticides, pp. 493–506.

Marcel Dekker, New York.

Hink WF & Strauss EM (1976) Growth of Trichoplusia ni (TN-368)

cells in suspension culture. In: Kurstak E & Maramorosch K (eds)

Invertebrate Tissue Culture, pp. 297–300. Academic Press, New

York.

Hink WF & Strauss EM (1980) Semi-continuous culture of the TN-

368 cell line in fermenters with virus production in harvested

cells. In: Kurstak E, Maramorosch K & Dübendorfer A (eds)

Invertebrate Systems In Vitro, pp 27–33. Elsevier-North Holland

Biomedical Press, Amserdam.

188

Jäger V (1996) Perfusion culture of insect cells. This volume.

Jain D, Ramasubramanyan K, Could S, Lenny A, Candelore M,

Tota M, Strader C, Alves K, Cuca G, Tung JS, Hunt G, Junker B,

Buckland BC & Silberklang M (1991) Large-scale recombinant

protein production using the insect cell-baculovirus expression

vector system: antistasin and receptor. In: Spier RE,

Griffiths JB & Meignier B (eds.) Production of Biologicals from

Animal Cells in Culture, pp. 345–350. Butterworth-Heinemann,

Oxford.

Junker BH, Hunt G, Burgess B, Aunins J & Buckland BC (1994)

Modified microbial fermenter performance in animal cell culture

and its implications for flexible fermenter design. Bioprocess

Eng. 11:57–63.

Kamen AA, Tom RL, Caron AW, Chavarie C, Massie B & Archam-

bault J (1991) Culture of insect cells in a helical ribbon impeller

bioreactor. Biotechnol. Bioeng. 38: 619–628.

Kamen AA & Tom R (1994) Mass spectroscopy determination of

insect cell respiration rates in culture and production process-

es. In: Spier RE, Griffiths JB & Berthold W (eds) Animal Cell

Technology. Products of Today, Prospects for Tomorrow (12th

Meeting of the European Society for Animal Cell Culture Tech-

nology) pp. 345–350. Butterworth-Heinemann, Oxford.

Kim K & Familetti PC (1994) Optimal culture conditions for the

production of recombinant proteins from baculovirus infected

cells. In Vitro 30: 722 (Presented at the 1994 Cell and Tissue

Culture Congress, Research Triangle Park, NC, 4-7 June 1994).

King GA, Daugulis AJ, Faulkner P, Bayly D & Goosen MFA (1988)

Growth of baculovirus-infected insect cells in microcapsules to a

high cell and virus density. Biotechnol. Lett 10: 683–688.

King G, Kuzio J, Daugulis A, Faulkner P, Alien B, Wu J & Goosen

MFA (1991) Assessment of virus production and chlorampheni-

col acetyl transferase expression by insect cells in serum-free and

serum-supplemented media using a temperature-sensitive bac-

ulovirus. Biotechnol. Bioeng. 38: 1091–1099.

King GA, Daugulis AJ, Faulkner P & Goosen MFA (1992) Recombi-

nant production in serum-free medium in a 14-L

airlift bioreactor. Biotechnol. Progr. 8: 567–571.

Kioukia N, Nienow AW, Emery AN & Al-Rubeai M (1995) Physi-

ological and environmental factors affecting the growth of insect

cells and infection with baculovirus. J. Biotechnol. 38: 243–251.

Klöppinger M, Fertig G, Fraune E & Miltenburger HG (1990) Mul-

tistage production

Autographa california nuclear polyhedrosis

virus in insect cell cultures. Cytotechnology 4: 271–278.

K16ppinger M, Fertig G, Fraune E & Miltenburger HG (1991) High

cell density perfusion culture of insect cells for production of

baculovirus and recombinant protein. In: Spier RE, Griffiths JB

& Meignier B (eds) Production of Biologicals from Animal Cells

in Culture (10th Meeting of the European Society for Animal

Cell Culture Technology) pp. 470–476. Butterworth-Heinemann,

Oxford.

Kompier R, Tramper J & Vlak JM (1988) A continuous process for

production of baculovirus using insect-cell cultures. Biotechnol.

Lett. 10: 849–854.

Kompier R, Kislev N, Segal I & Kadouri A (1991) Use of a sta-

tionary bed reactor and serum-free medium for the production of

recombinant protein in insect cells. Enzyme Microb. Technol. 13:

822–827.

Konstantinov KB, Pambayun R, Matanguihan R, Yoshida T, Perusich

CM & Hu W-S (1992) On-line monitoring of hybridoma cell

growth using a laser turbidity sensor. Biotechnol. Bioeng. 40:

1337–1342.

Kool M, Vondken JW, van Lier FLJ, Tramper J & Vlak JM (1991)

Detection and analysis

Autographa californica nuclear polyhe-

drosis virus mutants with defective interfering properties. Virol-

ogy 183: 739–746.

Lazarte EJ, Tossin P-F & Nicolau C (1992) Optimization of the

production of full-length rCD4 in baculovirus-infected Sf9 cells.

Biotechnol. Bioeng. 40: 214–217.

Licari P & Bailey JE (1991) Factors influencing recombinant pro-

tein yields in an insect cell – baculovirus system: multiplicity of

infection and intracellular protein degradation. Biotechnol. Bio-

eng. 37: 238–246.

Licari P & Bailey JE (1992) Modeling the population dynamics of

baculovirus-infected insect cells: optimizing infection stategies

for enhanced recombinant protein yields. Biotechnol. Bioeng. 39:

432–441.

Lindsay DA & Betenbaugh MJ (1992) Quantification of cell cul-

ture factors affecting recombinant protein yields in baculovirus-

infected insect cells. Biotechnol. Bioeng. 39: 614–618.

Maiorella B, Inlow D, Shauger A & Harano D (1988) Large-

scale insect cell culture for recombinant protein production.

Bio/Technol. 6: 1406–1410.

Miltenburger HG & David P (1980) Mass production of insect cells

in suspension. Devel. Biol. Standard. 46: 183–186.

Murhammer DW & Goochee CF (1988) Scaleup of insect cell cul-

tures: protective effects of Pluronic F-68. Bio/Technol. 6: 1411–

1418.

Neutra R, Levi B-Z & Shoham Y (1992) Optimization of protein

production by the baculovirus expression vector system in shake

flasks. Appl. Microbiol. Biotechnol. 37: 74–78.

Ogonah O, Shuler ML & Granados RR (1991) Protein production (

galactosidase) from a baculovirus vector in Spodoptera frugiper-

da and Trichoplusia ni cells in suspension culture. Biotechnol.

Lett. 13: 265–270.

Papoutsakis ET (1991) Media additives for protecting freely sus-

pended animal cells against agitation and aeration damage. Trends

Biotechnol. 9: 316–324.

Reuveny S, Kemp CW, Eppstein L & Shiloach J (1992) Carbohy-

drate metabolism in insect cell cultures during cell growth and

recombinant protein production. Ann. N.Y. Acad. Sci. 665: 220–

237.

Reuveny S, Kirn YJ, Kemp CW & Shiloach J (1993a) Production of

recombinant proteins in high-density insect cell cultures. Biotech-

nol. Bioeng. 42: 235–239.

Reuveny S, Kim YJ, Kemp CW & Shiloach J (1993b) Effect of

temperature and oxygen on cell growth and recombinant protein

production in insect cell cultures. Appl. Microbiol. Biotechnol.

38: 619–623.

Rhodes M, Gardiner S & Broad D (1991) Scaleup of animal cell

suspension culture. In: Ho CS and Wang DIC (eds.) Animal Cell

Bioreactors, pp. 253–268. Butterworth-Heinemann, Boston.

Schlaeger E-J, Foggetta M, Vonach JM & Christensen K (1993)

SF-1, a low cost culture medium for the production of recom-

binant proteins in baculovirus infected insect cells. Biotechnol.

Techniques 7: 183–188.

Schlaeger E-J, Wippler J & Foggetta M (1994) Scale-up of bac-

ulovirus expressed recombinant proteins using the High Five

insect cell line. Presented at Cell Culture Engineering IV, Engi-

neering Foundation Conferences, San Diego, CA 7-12 March

1994 (Abstract WE2, Cell Culture Engineering IV Program and

Abstracts, p. 85).

Schopf B, Howaldt MW & Bailey JE (1990) DNA distribution and

respiratory activity

Spodoptera frugiperda populations infect-

ed with wild-type and recombinant Autographa California nuclear

polyhedrosis virus. J. Biotechnol. 15: 169–186.

189

Scott RI, Blanchard JH & Ferguson CHR (1992) Effects of oxygen

on recombinant protein production by suspension cultures of Sf9

insect cells. Enzyme Microb. Technol. 14: 798–804.

Shuler ML, Cho T, Wickham T, Ogonah O, Kool M, Hammer DA,

Granados RR & Wood HA (1990) Bioreactor development for

production of viral pesticides or heterologous proteins in insect

cell cultures. Ann. N.Y. Acad. Sci. 589: 399–422.

Streett DA & Hink WF(1978) Oxygen consumption of TN 368 cell

line infected with AcNPV. J. Invertebr. Pathol. 31: 112–122.

Tramper J, Williams JB, Joustra D & Vlak JM (1986) Shear sensi-

tivity of insect cells in suspension. Enzyme Microb. Technol. 8:

33–36.

Tramper J, de Gooijer KD & Vlak JM (1993) Scale-up considera-

tions and bioreactor development for animal cell cultivation. In:

Goosen MFA, Daugulis AJ & Faulkner P (eds) Insect Cell Culture

Engineering, pp. 139–177. Marcel Dekker, New York.

van Lier FLJ, van den End EJ, de Gooijer CD, Vlak JM & Tramper

J (1990) Continuous production of baculovirus in a cascade of

insect-cell reactors. Appl. Microbiol. Biotechnol. 33: 43–47.

van Lier FLJ, van der Meijs WCJ, Grobben NA, Olie RA, Vlak JM

& Tramper J (1992) Continuous production with

a recombinant baculovirus insect-cell system in bioreactors. J.

Biotechnol. 22: 291–298.

Vaughn JL (1976) The production of nuclear polyhedrosis virus in

large-volume cell cultures. J. Invertebr. Pathol. 28: 233–237.

Wang M-Y, Kwong S & Bentley WE (1993a) Effects of oxy-

gen/glucose/glutamine feeding in insect cell baculovirus protein

expression: a study on epoxide hydrolase production. Biotechnol.

Progr. 9: 355–371.

Wang M-Y, Vakharia V & Bentley WE (1993b) Expresion of epoxide

hydrolase in insect cells: a focus on the infected cell. Biotechnol.

Bioeng. 42: 240–246.

Wang MY, Valler M, Pulliam TR & Bentley WE (1995) Elicitation of

protease activity in Spodoptera frugiperda cell cultures. Present-

ed at the 209th American Chemical Society National Meeting,

Anaheim, CA 2-6 April 1995 (Abstract BIOT 120).

Weiss SA & Vaughn JL (1986) Cell culture methods for large-scale

propagation of baculoviruses. In: Granados RR and Frederici BA

(eds) The Biology of Baculoviruses, Vol. 2 Practical Applications

for Insect Control, pp. 64–87. CRC Press, Boca Raton.

Weiss SA, Smith GC, Kalter SS & Vaughn JL (1981) Improved

method for the production of insect cell cultures in large volume.

In Vitro 17: 495–502.

Weiss SA, Orr T, Smith GC, Kalter SS, Vaughn JL & Dougherty

EM (1982) Quantitative measurement of oxygen consumption in

insect cell culture infected with polyhedrosis virus. Biotechnol.

Bioeng. 24: 1145–1154.

Weiss SA, DeGiovanni A, Godwin G & Kohler J (1988) Large scale

cultivation of insect cells. In: Roberts DW and Granados RR

(eds) Conference on Biotechnology Biological Pesticides and

Novel Plant-Pest Resistance for Insect Pest Management, Boyce

Thompson Institute for Plant Research, Cornell University, July

18-20, 1988, pp. 22–30 Cornell University Press, Ithaca, NY.

Weiss SA, Belisle BW, DeGiovanni A, Godwin G, Kohler J & Sum-

mers MD (1989) Continuous cell lines as substrates for biologi-

cals. Devel. Biol. Standard. 70: 272–279.

Wickham TJ, Davis T, Granados RR, Hammer DA. Shuler ML &

Wood HA (1991) Baculovirus defective interfering particles are

responsible for variations in recombinant protein production as

a function of multiplicity of infection. Biotechnol. Lett 13: 483–

488.

Wickham TJ, Davis T, Granados RR, Shuler ML & Wood HA (1992)

Screening of insect cell lines for the production of recombinant

proteins and infectious virus in the baculovirus expression sys-

tem. Biotechnol. Progress 8: 391–396.

Wickham TJ & Nemerow GR (1993) Optimization of growth

methods and recombinant protein production in BTI-TN-5B1-4

insect cells using the baculovirus expression vector. Biotechnol.

Progress 9: 25–30.

Wong TKK, Nielsen LK, Greenfield PF & Reid S (1994) Relation-

ship between oxygen uptake rate and time of infection of Sf9

insect cells infected with a recombinant baculovirus. Cytotech-

nology 15: 157–167.

Wood HA, Johnston LB & Burand JP (1982) Inhibition of Auto-

grapha californica nuclear polyhedrosis virus replication in high-

density Trichoplusia ni cell cultures. Virology 119: 245–254.

Wu J, King G, Daugulis AJ, Faulkner P, Bone DH & Goosen, MFA

(1989) Engineering aspects of insect cell suspension culture: a

review. Appl. Microbiol. Biotechnol. 32: 249–255.

Wu JY & Goosen MFA (1996) Insect cell immobilization. This

volume.

Zhang J, Kalogerakis N, Behie LA & latrou K (1993) A two-stage

bioreactor system for the production of recombinant proteins

using a genetically engineered baculovirus/insect cell system.

Biotechnol. Bioeng. 42: 357–366.

Zhang J, Kalogerakis N & Behie LA (1994a) Optimization of the

physicochemical parameters for the culture of Bombyx mori insect

cells used in recombinant protein production. J. Biotechnol. 33:

249–258.

Zhang J, Kalogerakis N, Behie LA & latrou K (1994b) Optimum

infection conditions for recombinant protein production in insect

cell (Bm5) suspension culture. Biotechnol. Progr. 10: 636–643.

Zhou W & Hu W-S (1994) On-line characterization of a hybridoma

cell culture. Biotechnol. Bioeng. 44: 170–177.

Address for correspondence

S.N. Agathos, Unit of Bioengineering,

Catholic Univeersity of Louvain, Place Croix du Sud 2/19, B-1348

Louvain-la-Neuve, Belgium.

!"#$%&'()%#*+)*+#,*'--.%-)/+%0-'*1

Cytotechnology 20: 191–198, 1996.

191

Perfusion bioreactors for the production of recombinant proteins in insect

cells

Volker Jäger

Abteilung Zellkulturtechnik, Gesellschaft für Biotechnologische Forschung GmbH, Mascheroder Weg 1, D–38124

Braunschweig, Germany

Key words: insect cell culture, perfusion culture, membrane perfusion, crossflow microfiltration, baculovirus,

bioreactor, fluidized bed, packed bed, recombinant protein production

Introduction

One of the fundamental reasons for using perfusion for

insect cell culture is to reach higher cell densities which

result in higher product titers and volumetric productiv-

ities. The growth of cells in customary insect cell cul-

ture media is usually limited to a maximum cell density

of However, cell densities of

more than can be obtained by opti-

mizing these media formulations by simple addition of

a few nurients such as glucose and glutamine (Acker-

mann et al., 1994). Cell densities of up to

were achieved using more sophisticated feeding

strategies (Bédard et al., 1995). In contrast to mam-

malian cells there appears to be no critical accumula-

tion of toxic metabolic products at these cell densities.

Therefore, one of the most important advantages of

continuous perfusion culture, namely the removal of

cytotoxic metabolites such as ammonium, appears to

be irrelevant to insect cell culture.

Typical problems of perfusion culture include the

more complex bioreactor setup and in the case of cells

requiring Baculovirus infection with baculovirus an

increasingly unfavourable ratio of virus inoculum to

the total reactor volume. Furthermore, the cell specific

productivity of Trichoplusia ni cells can be drastical-

ly reduced at higher cell densities. This ‘cell density

effect’ has been observed with Tn 368 cells (Wood

et al., 1982) as well as with BTI-Tn 5B1–4 cells

(Chung & Shuler, 1993). Another problem is the rapid-

ly declining production of infectious virus particles

with increasing number of passages, the so-called ‘pas-

sage effect’ which is encountered during continuous

production processes (Kompier et al., 1988; Tram-

per et al., 1990; van Lier et al., 1990). As expected,

recombinant protein production is also reduced by this

effect (van Lier et al., 1992, 1994). However, when

the production of large amounts of recombinant pro-

tein requires multiple virus passages this phenomenon

occurs regardless of the use of perfusion culture or a

cascade of batch reactors.

On the other hand, the use of perfusion culture in

insect cell cultivation and production of recombinant

proteins features numerous advantages. In addition to

the possible increase in cell specific and volumetric

productivity, the residence time of secreted recombi-

nant proteins in the bioreactor is significantly reduced,

thus minimizing product degradation. By using micro-

filtration systems, the first step of downstream process-

ing, the separation of a secreted recombinant protein

from cell debris is already integrated into the produc-

tion process. Furthermore, because of the continuous

supply of culture medium which prevents cell death

due to lack of nutrients, the production process is virtu-

ally independent of the speed of infection. This enables

the use of an extremely low multiplicity of infection

for the production of secreted proteins.

Internal membrane perfusion

Continuous medium exchange by internal membrane

perfusion does not require additional pumping of cells

through a filtration system. Therefore, cells are not