12_ANIMAL CELL TECHNOLOGY_707

605

the structural integrity of at least one sialic acid molecule. These findings would fit with the
hypothesis that gangliosides act by inducing conformational changes in the CD4 molecule
[5] prior to its internalization and degradation.
Due to their probable potential to pass the plasma membrane, the strong effects of GM3
derivatives D2, D3, D4 and D5 (see Fig. 1) lacking long chain fatty acids suggest the
existence of an alternative until yet unknown mechanism of CD4 down-modulation.

4. References

[ I ] Igarashi, Y., Nojiri, H., Hanai, N., and Hakomori, S.-I. (1989) Gangliosides that modulate membrane
protein function, Methods Enzymol. 179, 521-540.

[2] Hakomori, S.-I. (1990) Bifunctional role of glycosphingolipids, J. Biol. Chem. 265, 18713-18716.
[3] Zeller, C.B., and Marchase, R.B. (1992) Gangliosides as modulators of cell function, Am. J. Physiol. 262,

C1341-C1355.
[4] Sattentau, Q. J., and Weiss, R. A. (1988) The CD4 antigen: physiological ligand and HIV receptor, Cell

52, 631-633.
[5] Saggioro, D., Sorio, C., Calderazzo, F., Callegaro, L., Panozzo, M., Bertons, G., and Chieco-Bianchi, L.

(1993) Mechanism of action of the monosialoganglioside GM1 as a modulator of CD4 expression, J. Biol.
Chem. 268, 1368-1375.
[6] Ledeen, R. W., and Yu, R. K. (1982) Gangliosides: structure, isolation and analysis, Methods Enzymol..
83, 139-191.
[7] Müthing, J., and Unland, F. (1994) Improved separation of isomeric gangliosides by anion-exchange high-
performance liquid chromatography, J. Chromatogr. 658, 39-45.
[8] Lanne, B., Uggla, L., Stenhagen, G., and Karlsson, K.-A. (1995) Enhanced binding of enterotoxigenic
Escherichia coli K99 to amide derivatives of the receptor ganglioside NeuGc-GM3, Biochemistry 34,
1845-1850.
[9] Nores, G. A., Hanai, N., Levery, S. B., Eaton, H. L., Salyan, M. E. K., and Hakomori, S.-I. (1989)
Synthesis and characterization of ganglioside GM3 derivatives, Methods Enzymol. 179, 242-253.
[10] Ciucanu, I., and Kerek, F. (1984) Rapid and simultaneous methylation of fatty acids and hydroxy fatty
acids for gas-liquid chromatographic analysis, Carbohydr. Res. 131, 209-218.
[11] J., and Egge, H. (1990) Desorption mass spectrometry of glycosphingolipids, Methods
Enzymol. 193, 713-733.

5. Acknowledgements

This work was financed by the Deutsche Forschungsgemeinschaft SFB 223. The participation on the 15th
ESACT Meeting was supported by the European Commission Directorat General XII.

HUMAN SERUM IN LEUKOCYTE CULTURES PRODUCING HUMAN
INTERFERON ALPHA

P.Mattana, L. Scapol, S. Silvestri, and G. C. Viscomi.

Biotech.&Immunol. Dept., Alfa Wassermann SpA, Via Ragazzi del ‘99, 5
I-40133 Bologna, Italy

Keywords: leukocyte alpha interferon, human agamma serum.

ABSTRACT
Interferon Alpha (IFNα) is an antiviral, antiproliferative and immunostimulant cytokine
world-wide applied in the treatment of several viral and oncologic diseases. Among the
different preparations of , namely leukocyte lymphoblastoid
and recombinant LE- is particularly relevant, since it is better
tolerated and it does not induce any antigenic reaction to the administred drug. LE-
is produced through the induction by viruses, such as Sendai virus, of human
purified leukocyte cultures, whose media contain human agamma serum (HAS). Any
attempt to omit the use of HAS caused poor yield of LE- Therefore, the role of
HAS was investigated testing different preparations, performing dose (HAS
concentration)/response curves, and evaluating the importance of the
presence of HAS during the time of the culture. No differences in production
were detected based on the different HAS preparations. It was evident the dependence
of LE- yield on the HAS concentration, even if some differences could be detected
between different preparations. Optimal amount of HAS was depended on the cell
concentrations. Reduction of HAS concentration was possible only after having started

the culture at the optimized conditions. It could be concluded that an optimal
concentration of HAS per cell can be defined and the presence of HAS is critical in
particular at the beginning of leukocyte cultures, even before the addition of Sendai
virus, suggesting that the main role of HSA could be to restore the capacity of purified
leukocytes to produce rather than to support the expression during the

culture.

INTRODUCTION a complex family of several proteins, is one of the most
Alpha interferon
physiologically important defenses of leukocytes against foreign agents such as viruses
and Several studies have been published indicating that pharmaceutical
preparations, i.e. recombinant, lymphoblastoid and leukocyte interferon, LY-
and respectively) offer therapeutic benefits in patients with viral
diseases, hematologic and solid tumors.(1,2) Among the
pharmaceutical
preparations '. does not induce antibodies to during therapy,
since it is the most similar to the physiologically expressed while the other
do. Furthermore, is effective in restoring response in patients who lost

response following production of

607

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 607-612.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

608

The genes are transiently expressed only under appropriate stimuli, such as
the addition of viruses, by primary leukocytes harvested from healthy donor blood
donations and cultured in a Minimum Essential Medium (MEM) supplemented with
gamma globulin-free human serum (human agamma serum, HAS). After six hours from
infection reach a maximun steady state and decline in the remaining
time of the culture, and in the meantime proteins are released in the culture
supernatants.(3 ,4)
The optimization of leukocyte culture parameters is critical to maximize the
yields, since even minute changes could have significant effects. In this respect the role
of HAS was investigated by testing different preparations, performing dose (HAS
concentration)/response curves, and checking the HAS concentration with
respect to the age and concentration of the leukocytes, and the culture time.

MATERIALS AND METHODS
HAS preparation - The tested HAS preparations were prepared by combining described
procedures.(5 ,6) They were prepared starting from different plasma sources: fresh frozen
(FF) plasma, plasma harvested from buffy coat pools and plasma collected from single
buffy coats. In all the cases the sera were obtained by freezing/thawing the plasma or by
adding 0.7 mg/ml of calcium chloride to the plasmas at 4°C over night. The removal of

gamma globulins from the sera was obtained by precipitation at 4°C over night with
polyethylenglycol 6000 (PEG) at the concentration of 6% (w/v) or with 40% (v/v) of
saturated ammonium sulfate. The precipitated gamma globulins were removed by
centrifugation and when ammonium sulfate was used the solutions were also dialyzed.
All the HAS preparations, after sterile filtration, were inactivated at 56°C for 30 min,
stored at -20° C and thawed at 4° C just before use.
The HAS preparations were tested on the basis of
production following the Cantell’s procedure.(5) Human leukocytes, obtained
from 24-hour-old or 48-hour-old buffy coats (BCs) after lysis of erythrocytes by 0.83%
ammonium chloride treatment, were cultured at two different cell concentrations: 4x106
and 10x106 cells/ml at 37°C in a MEM containing HAS of different preparations and at
the concentrations ranging from 0.2 to 2.2 mg/ml, according to what is specified in each
experiment.
The leukocyte cultures (each of 100 ml in round-bottom flask) were induced to produce
by adding 100-200 IU/ml of as primer and after 2 hours by adding
150 haemmaglutinating units per milliliter (HAU/ml) of Sendai virus. After 4.5 hours
the temperature was lowered to 30° C. The cultures were stopped after 18-20 hours and
the produced was collected by centrifugation. Leukocyte cultures were also
performed in the absence of HAS and by replacing HAS with 2.0 mg/ml of human
serum albumin. During the culture time, changes in the cell concentration alone or in
combination with modifications of HAS concentration were carried out by properly
diluting the cultures with prewarmed MEM with and without added HAS. Changes in
HAS concentration alone were performed by centrifuging the leukocyte cultures and
resuspending the pellets in media with the new HAS concentration to be tested.
concentrations were determined by evaluating the antiviral activity through the
reduction of the cytopathic effect assay.(7 ,8)
Statistical Analysis - The Student’s t test was applied. A p value was considered
significant.

609

RESULTS
The effect on the yields of the different HAS preparations are provided in
Table 1, where the specific preparation procedures of each HAS lot can be deduced
from inspection of column 1-5.

The reported yields show differences, which, however, are not statistically significant,
therefore the remaining part of the experimentation was performed using only HAS-1
and HAS-2, which provided the highest titers.
These two HAS lots, prepared with the same procedure, were analyzed by performing a
dose/response experiment. The experimental points were fitted with the function

where y represents concentration; x the HAS concentration; a,b,c, and d are
constants (Figure 1). The results indicate that the interpolated curves of HAS-1 and
HAS-2 had similar shapes; nevertheless, since the plateaus were reached at different
concentrations, 1.5 and 3.53 mg/ml, respectively, the curves did not overlap in the
range from 0.5 to 1.8 mg/ml of HAS.

Based on these curves, the concentrations of 1.5 and 2.0 mg/ml for HAS-1 and HAS-2,
respectively, were chosen among those possible, since they are the closest values which
allow . yields higher than 90% of the plateaus for both the HAS preparation.
When the addition of HAS to the leukocyte cultures was omitted and when HAS was
replaced by identical amounts of human albumin, significant reduction of the
yields, and respectively, was obtained and these values

610

correspond to approximately 20% of the productivity in the presence of optimal

amounts of HAS. yields of induced leukocyte
The figure 2 shows the examination of the
cultures, coming from 24-hour-old BCs (Figure 2A) and 48-hour-old BCs (Figure 2B)
performed at and at cells/ml, in which HAS-2 was used at the
concentration of 2.0 mg/ml and at 0.8 mg/ml, where the later value was obtained by
reducing 2.0 mg/ml by the same factor used to diluted the cell concentration from

In the case of 24-hour-old BCs the results indicate that by keeping the ratio of HAS
amount per cell constant at 0.2 ng/cell or increasing it at 0.5 ng/cell, a similar
productivity of per cell can be achieved even at different absolute values of
HAS. When this ratio was lowered to 0.08 ng/cell, as in the case of cells/ml
cultures at 0.8 mg/ml, statistically significant reduction of yields were
obtained (Figure 2A). A different behaviour in 48-hour-old leukocyte cultures can be
appreciated, since at the ratio of HAS/cell of 0.5 ng/cell at the cell concentration of
cells/ml a significant increase of yield respect to those obtained with 0.2
ng/cell both at and cells/ml was reached (Figure 2B).

When cells/ml cultures, whose leukocytes were obtained from 24- and 48-hour-
old BCs, were diluted down to cells/ml, after 4.5 hour from the beginning, and the
HAS concentration was either maintained (1.5 mg/ml, since HAS-1 was used) or
reduced proportionally to the cell concentration (0.6 mg/ml), the obtained yields did not
show any significant differences, even when compared with the productivity

611

of a culture, whose HAS-1 and cell concentrations (1,5 mg/ml and cells/ml,
respectively) were kept constant (Table 2).
Reductions of HAS concentration under the optimal value of 2.0 mg/ml were
performed during the culture to assess the importance of HAS concentrations in relation
to culture time (Table 3).

As controls of the conditions 2,3 and 4 leukocyte cultures were centrifuged at the same
time and resuspended in the initial HAS concentration. The following yields were
obtained: 9,268 ± 2,352, 8,765 ± 2,964 and 8,125 ± 1,812 IU/million cells,
respectively.
The analysis of the results reveals that respect to the culture kept at the suboptimal
HAS concentration for all the time differences in the yields are detectable
even in those leukocyte cultures whose HAS concentration was at the optimal value for
only 1 hour, even before the addiction of the stimulus to induce .. However the
differences became significant after 3 hours.

DISCUSSION

synthesis by primary leukocytes under viral stimuli is a transient
event, which does not take more than 6 hours from infection. After this point a down-
regulation mechanism takes place which strongly reduces any further
synthesis.(3,9) Therefore, since no leukocyte growth is expected in the described culture
conditions, the production can be considered as a unique and irreversible
event. In vitro synthesis does not require any specific medium component,
since has been also produced in absence of sera, amino acids and vitamins.
(5,10) Nevertheless, when an optimization process is carried out to maximize the LE-
production by human leukocytes, several components of the culture medium
became important. In the present work the contribution of HAS to the improvement of
the yield was studied.
Even though the specific HAS components which play a relevant role in
production are unknown, it can still be concluded from the data of Table 1 that the
choice of the precipitation agent, PEG or ammonium sulfate, and that the source of the
plasma, fresh-frozen plasma, plasma from BC pools or plasma from single BCs, did not
introduce any relevant differences in those HAS components, which could be essential
for the production. While some of the differeces detected from lot to lot
should be attributed to the quality of the plasma (donors, anticoagulants, storage
conditions, transport) used for the HAS preparation.

612

This consideration was confirmed by the dose/response curves in Figure 1, where the
yields are correlated with the quality of HAS preparations obtained by
applying the same preparation procedure, thus allowing for each HAS lot the definition
of the concentrations at which production is maximized at the minimal
protein concentration in the leukocyte cultures.
The main role of HAS in improving the yield cannot be attributed either to
an enhancement of stability or to a protection of from aspecific
proteolysis, since the yields obtained with human albumin in place of HAS were
similar to those obtained in absence of any protein stabilizer.
The optimal HAS concentration is related to the cell concentration and to the leukocyte
age. It could be proportionally reduced, if more diluted cells are used and it has to be
optimized on the basis of BC age, since leukocyte cultures coming from 48-hour-old
BCs seems to require higher HAS concentration than that of 24-hour-old BC to
maximize the yields (Figure 2 and Table 2).
The importance of the use of HAS at the optimal concentration seems to be restricted to
the early hours of the culture (Table 3). This result is conceivable, since, as was
mentioned before, the mRNA expression in leukocytes is a fast event, which
reaches the maximun in only six hours.
In conclusion, production by human leukocytes is a unique and transient

event, which is affected by even minute changes of the culture conditions. Among them
the concentration and the quality of HAS, the age of BCs, and the cell concentration
resulted to be important. From the data reported, it could stated that an optimal
concentration of HAS amount per cell can be defined to maximize
production depending on the HAS preparation and the age of BCs. Furthermore, the
presence of HAS is critical at the beginning of leukocyte cultures, even before the
addition of any stimulus to produce suggesting that the main role of HAS
could be to restore the capacity of purified leukocytes to produce rather than to
support the expression during the culture.

REFERENCES

1 In Interferon, Priciples and Medical Applications, Baron, S., Coppenhaver, D.H., Dianzani, F.,
Fleishmmann, Jr, W.R., Hughes, Jr, T.K., Kimpel, G.R., Niesel, D.W., Stanton, G.J. and Tyring, S.K. Eds.,
The University of Texas Medical Branch of Galveston, Department of Microbiology, Galveston , TX, 1992.
2 Viscomi, G.C., Grimaldi, M., Palazzini, E., and Silvestri, S. (1995) Human leukocyte interferon alpha:
structure, pharmacology, and therapeutic application, Medicinal Research Reviews 15, 445-478.
3Hiscott, J., Cantell, K., Weissmann, C. (1984) Differential expression of human interferon genes, Nucleic.
Acids. Res. 12, 3727-3746.
4Pitha, P.M., Au, W.C. (1995) Iduction of interferon alpha genes expression, Semin. Virol 6, 151-159.
5 Cantell, K., H i r v o n e n , S., Kauppinen, K.L., Myllyla, G. (1981)Production of interferon in human
leukocytes from normal donors with the use of sendai virus, Methods in Enzymol. 78, 29-38.
6 Macleod, A.J. (1991) Serum and its fractionation. In M.Butler (ed.), Mammalian Cell Biotechnology, IRL
Press New York, pp 27-37.
7 Armstrong, J.A. (1981) Cytopatic effect inhibition assay for interferon: microculture plaque assay, Methods
in Enzymol. 78, 381-387.
8 Berg, K., Hansen, M.B., Nielsen, S.E. (1990) A new sensitive bioassay for precise quantification of
interferon activity as measured via the mitocondrial deydrogenase function in cells (MTT-METHODS),
AMPIS. 98, 156-162
9 Tovell, D., Cantell, K . , (1971) Kinetics of interferon production in human leukocyte suspensions,
J.Gen.Virol. 13, 484-489.
10 Goore, M.Y, Dickson, J.H., Lipkin, S., and DiCuollo, C.J. (1973). The production of Human Leukocyte
Interferon in a Serum-free Medium, Proc.Soc.Exp.Biol. Med. 142, 46-49.

THE NEW TYPE OF IMMUNOMODULATOR

M.V.MEZENTSEVA*, V.A.MOZGOVOI,
L.Yu.MOZGOVAYA, R.Ya.PODCHERNYAEVA.
*N.F.Gamaleya Institute of Epidemiology and Microbiology,
D.I.Ivanovsky Institute of Virolology
Moscow, Russia

1. Introduction

There is quite a number of substances (both synthetic or of plant origin) capable of modulating the
immune system of human beings and animals, either enhancing, or decreasing the interferons
(IFN) and other cytokines production [3]. All of them are introduced s/c, i.v., i.p., per os or per
rectum. Interferons of different types are synthesized by different categories of
immune competent cells. Earlier studies using monoclonal antibodies revealed that lymphocytes
and macrophages produced mostly type I IFN, while T-cells produced type II IFN [2].
It was exciting to study the production of endogenous IFN by normal murine cells belonging to
different populations following Applicator usage. The ability to produce IFN may be used for
assessment of the immune cell functional activity [5].

2. Materials and methods

Applicator [4] is a disk (45-50 mm in diameter, 7 mm thick), made out of the biopolimer with
biologically active metal suspension (Cu, Fe, Zn, Al, Co, Ni, Na, Li, Mg) with variable valency in
the ionized high-spin condition. The structure of the metallic ingredients has similarity with that of
the microelements in the active center of proteins.
A study was made using CBA mice and hybrids. Applicator was
applied from the distance of 15 cm from mice or in the direct contact. Time of application varied
from 15 to 40 min daily during the 5-day period. The murine immune cells were isolated from
thymus (mostly T-cells), bone marrow (mostly B-cells), and spleen - cell mixture, 30% of which are
active macrophages, and the remaining 70% - lymhocytes, from which T-cells comprise 60%, and
B-cells - 40% [2].
IFN litres were estimated in mice blood and immune cells using standard procedure [1, 2].

3. Results and discussion

The IFN-inducing ability of Applicator was studied in mice during the 5-day period. Serum IFN
titers were estimated daily. Fig. 1 shows that Applicator usage enhanced the IFN production 5-32-
fold.

613

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 613-615.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

614

The main cells producing IFN both in men and animals are cells of the immune system.
Substances of different classes are able to stimulate the production of antigenically different IFNs
in different kinds of cells [2]. As the IFN producing locally is known to play an important role in
the pathology of different infections, it was interesting to study the influence of Applicator on the
IFN production in the thymus, bone marrow and spleen cells.
Fig.2 depicts the ability of Applicator to modulate the IFN production in thymus cells,

mostly consisting of T-cells, following its application. Usually thymocytes produce while
synthesis of does not occur. But we managed to show that Applicator usage oppressed the
synthesis of by the thymus cells, while was produced by them at a high titer.

Bone marrow cells, consisting mostly of B-lymphocytes, in contrast to thymocytes, in

response to various inducers stimulation normally produce We shown that Applicator

usage completely suppressed the synthesis (Fig.3).

615

The production of in splenic cells (mixture of T-cells, B-cells and macrophages) is
shown in Fig.4. The splenocytes of mice that did not have a contact with Applicator synthesized

in the quantuty 80 IU/ml. Following the contact with Applicator the ability of splenocytes

to produce increased 16-fold.

The data shown above prove that Applicator used possess a high IFN-inducing activiy. In

normal mice it is shown to induce the production of endogenous Meanwhile Applicator is
probably able to modify the immune competent cells in the unusual way, somehow turning on
their concealed reserves.

4. Conclusions

1. Applicator usage is shown to induce the IFN-synthesis in the murine organism in the quantities
high enough to control different infectious pathology.
2. Applicator is probably able to modify the immune competent cells of mice in the unusual way,
somehow turning on their concealed reserves.
3. Applicator possess high IFN-inducing activity. It is shown to enhance the IFN synthesis up to
the titer of 320 IU/ml.
4. Applicator might be used for therapeutical aims, as it used outwardly. The course of the
treatment is 1-10 days.

5. References

1. Campbell, I., Greenber, T., Kochman, M.A. (1975) A microplaque reduction assay for
human and mouse interferon, Can. J. Microbiol. 21, 1247 - 1253.
2. Ershov, F.I. (1996) The Interferon System in Normal State and in Case of Pathology,

Medicine, Moscow.
3. Ershov, F.I., Chizhov, N.P., Tazulakhova, E.B. (1993) Antiviral medicines, Ingoda, S-

Peterburg.

4. Mozgovoi, V.A. and Mozgovaya, L.Yu. (1995) Applicator, Patent N 2040248, Russia.
5. Stewart II, W.E. (1979) The Interferon System, Springer-Verlag, New York.

IN VITRO IMMUNIZATION OF HUMAN PERIPHERAL
BLOOD LYMPHOCYTES WITH CHOLERA TOXIN B
SUBUNIT

A. ICHIKAWA, Y. KATAKURA, T. KAWAHARA, S. HASHIZUME*
and S. SHIRAHATA
Graduate School of Genetic Resources Technology, Kyushu University,
6-10-1 Hakozaki, Higashi-ku, Fukuoka 812, Japan
*Morinaga Institute of Biological Science, 2-2-1 Shimosueyoshi,
Tsurumi-ku, Yokohama 230, Japan

Abstract In vitro immunization (IVI) techniques have a great potential in the production
of human monoclonal antibodies (MAbs) against various antigens. An IVI method of
human peripheral blood lymphocytes (PBL) has been developed with a human lung
adenocarcinoma cell line in our laboratory. We have succeeded in generating several cancer
specific human MAbs using this IVI method. Because this IVI method was not available for
soluble antigens, we improved it for soluble antigens. IVI with soluble antigens was
effectively caused by the addition of muramyl dipeptides, interleukin-2 (IL-2) and
interleukin-4 (IL-4). It was found that the difference of sensitivity of lymphocytes
depending upon donors could be overcome to find the optimal concentrations of IL-2 and
IL-4. IVI of human PBL was performed with cholera toxin B subunit (CTB) and the
immunized B cells were transformed by Epstein-Barr virus. Anti-CTB antibody was
detected using an indirect ELISA. Clones producing anti-CTB antibodies were directly
cloned by a soft agar cloning method.

1. Introduction

Human monoclonal antibodies are useful for diagnosis and treatment of various diseases.
However, in order to obtain the human monoclonal antibodies, passive immunization,
such as injection of dangerous antigens in human bodies, should not be permitted for ethical
and moral reasons. One of the mothods to solve this problem is the in vitro immunization
method which is a primary activation of antigen-specific B lymphocytes. In vitro
immunization techniques have great potential in the production of human monoclonal
antibodies against various antigens. We have developed an in vitro immunization method
against cultured lung adenocarcinoma cell line1) and succeeded in generating several cancer
specific humanmonoclonal antibodies2), 3). However,this methodcouldnotbeavailablefor
immunization against soluble antigens. The purpose of this work is to improve this in
vitro immunization method for immunization against soluble antigens and to generate
human monoclonal antibodies against cholera toxin B subunit with new in vitro
immunization method.

617

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 617-623.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

618

2. Materials and Methods

2.1. Antigens and reagent

Keyhole lympet hemocyanin (KLH) was obtained from Calbiochem-Novabiochem
Corporation (LaJolla, CA, USA). Ovalbumin was obtained from Chemicon international
Inc. (Temecula, CA, USA). Cholera toxin B subunit was obtained from List Biological
Laboratories, Inc. (Campbell, CA, USA). Recombinant human IL-2, IL-6 and IL-10 were
purchased from Genzyme corporation (Cambridge, MA, USA). Recombinant human IL-
4 was purchased from Pepro Tech EC LTD. (London, England). Recombinant human IL-5
was obtained from R&D systems (Minneapolis, MN, USA). Muramyl dipeptide (MDP)
was purchased from Chemicon international Inc. (Temecula, CA, USA). L-Leucyl-leucin
methyl ester was obtained from Boehringer GmbH (Mannheim, Germany).

2.2. Isolation of human lymphocytes

Human peripheral blood lymphocytes (PBL) were separated by density-gradient
centrifugation from several healthy donors. In brief, 25 ml of peripheral blood was layered
on 20 ml of lymphocyte separation medium (LSM; Organon Teknika, Durham, NC, USA)

and was centrifugcd at 400g for 30 min. The stratum of lymphocytes were harvested and
washed three times with ERDF medium. Lymphocytes from peripheral blood were treated

with 0.25 mM Leu-Leu-OMe to remove the cytotoxic T cells, CD8+ suppressor T cells, and
natural killer cells before use.

2.3. In vitro immunization

In vitro immunization of the human PBL was performed in 24 well culture plates (Becton

Dickinson). The Leu-Leu-OMe-treated PBL were cultured for 7 days in ERDF medium
containing 10% heat inactivated fetal bovine serum, MDP (10µg/ml), IL-2 (10 units/ml),

IL-4 (10 ng/ml), 2-mercaptoethanol (20 µM) and antigens.

2.4. Enzyme-linked immuno sorbent assay

Ninety-six well microtiter plates were coated with CTB or BSA in 0.1 M Na-carbonate

buffer (pH9.6) at concentration of 0.2 µg/well and incubated at 4°C overnight. The plate

were washed three times with PBS/0.05% Tween 20 (PBS-T), the supernatants of EBV
transformed B cells were diluted 1:2 in PBS-T, added in , aliquots per well, and
incubated at 4°C overnight. The wells washed three times with PBS-T and incubated for 2
hours at 37°C with l/well horseradish peroxidase-conjugated goat antibodies against
human IgM. After washing three times with PBS-T, l/well of substrate solution [0.1

M citrate buffer (pH4.0) containing 0.003% and 0.3 mg/ml p-2.2’-azino-di (3-
ethylbenzothiazoline-6-sulfonic acid) diammonium salt] were added, and after 20 minutes,
the absorbance at 405 nm was measured by using an ELISA reader.

2.5. EBV transformation
After in vitro immunization, human B cell were Epstein-Barr virus (EBV) infected using

619
supenatants from the B95-8 marmoset cell line. EBV-transformedcells were washed with
ERDF medium containing 20% FBS and seeded into 96 well culture plates at
cells/well.
2.6. Cloning
EBV-B cells producing antibodies to choleratoxin B subunit were cloned by limiting

dilution at density of 1 cell/well in 96 well culture plates or by soft agarose colony forming
method. The supenatants from the positive wells were screened by ELISA.

3. Results and Discussion
3.1. Establishment of protocol for in vitro immunization with soluble
antigens

We examined the effect of various recombinant human cytokines which are concerned with
the proliferation and differentiation of B lymphocytes. Peripheral blood lymphocytes (PBL)
derived from healthy donors were incubated with various combinations of IL-2, IL-4, IL-
5, IL-6 and IL-10 in the presence of KLH, as an antigen and MDP for 7 days. Table 1
summarizes the effects of various cytokines on in vitro immunization against soluble

antigen. It was found that IgM production and cell proliferation were greatly increased in the
combination of IL-2 and IL-4. Similarly, we could observe high production of IgM reactive
to KLH. These results indicated that the protocol using the combination of IL-2 and IL-4
was the most effective for in vitro immunization against soluble antigens.

We established the new protocol of in vitro immunization for soluble antigens. Fig. 1
shows a scheme of the new protocol. Major difference of the protocol for soluble antigens
from that for cancer cell antigens was the use of IL-4 instead of IL-6 in the culture medium
for in vitro immunization of lymphocytes.

620

3.2. In vitro immunization with cholera toxin B subunit (CTB)
To illustrate the effect of the new protocol, we actually performed in vitro immunization of
human PBL with CTB as an example of soluble antigens. Fig. 2 shows reactivity of
antibodies induced by in vitro immunization of PBL with CTB as an antigen. We examined
the effects of several combinations of IL-2, IL-4 and IL-6 on in vitro immunization with
CTB. Although there were not much differences on immunoglobulin secretions among
every combination, the combination of IL-2 and IL-4 was most effective on the level of
CTB-specific antibodies. Exogenous addition of IL-6, which was effective for in vitro

621
immunization with cultured cell antigen, had no effect on immunization with soluble
antigens.

3.3. Individual differences of lymphocytes on in vitro immunization
All examinations mentioned above were carried out using PBL derived from single donor.

We examined whether this protocol was also effective for lymphocytes derived from
different donors or not. Fig. 3 shows results of in vitro immunization of lymphocytes
derived from three different donors. We could detect the increase of antigen specific IgM
antibody secreted by lymphocytes derived from donor A, but not on donor B and C. This
result seems to suggest the possibility that each lymphocytes derived from different donors

demand their own optimum concentrations of IL-2 and IL-4.

Then we tried to optimize the concentrations of IL-2 and IL-4 on in vitro immunization of
lymphocytes derived from different donors. Table 2 shows the optimum concentrations of
IL-2 and IL-4 in the culture of in vitro immunization of lymphocytes derived from different
donors. It was found that each lymphocytes from different donors demanded their own
optimum concentrations of IL-2 and IL-4. For this reason, if we can find optimum
concentrations of IL-2 and IL-4 for each lymphocytes beforehand, the effective
immunization will be caused by this protocol using the combination of IL-2 and IL-4.

622
3.4. Acquisition of EBV-B cell lines secreting CTB-specific MAbs
To immortalize the B lymphocytes after in vitro immunization with CTB, we adopted the
Epstein-Barr virus (EBV) transformation. Fig. 4 shows screening of EBV-B lymphocytes
supernatants in ELISA. Several EBV-B lymphocytes secreted antibodies reacting strongly
with CTB. But most supernatants also reacted to BSA as a control. Lymphocytes cultured
in the well no. 13C had the highest specificity to CTB of all the wells. Positive EBV-B
lymphocytes in 13C were amplified and cloned by soft agar cloning method. We could
obtain 4 clones secreting IgM antibodies reactive to CTB among 280 clones picked up.
However, only one clone, 13C/C7, had a specificity to CTB (Fig. 5).

4. Acknowledgment
This work is supported by a Grant-in-Aid for Scientific Research from the Ministry of
Education, Science and Culture of Japan. A. I. is a research fellow of the Japan Society
for the Promotion of Science.
5. References

1. Kawahara, H., Shirahata, S., Tachibana, H. and Murakami, H. (1992) In vitro
immunization of human lymphocytes with human lung cancer cell line A549. Hum.
Antibod. Hybridomas 3 , 8-13.

2. Shoji, M., Kawamoto, S., Sato, S., Kamei, M., Kato, M., Hashizume, S., Seki, K.,
Yasumoto, K., Nagashima, A., Nakanishi, H., Suzuki, T., Imai, T., Nomoto, K. and
Murakami, H. (1994) Specific reactivity of human monoclonal antibody AE6F4 against
cancer cells in tissues and sputa from lung cancer patients. Hum. Antibod. Hybridomas 5 ,
116-122.

3. Shoji, M., Kawamoto, S., Setoguchi, Y., Mochizuki, K., Honjoh, T., Kato, M.,
Hashizume, S., Hanagiri, T., Yoshimatsu, T., Nakanishi, K., Yasumoto, K., Nagashima,
A., Nakahashi, H., Suzuki, T., Imai, T., Nomoto, K. and Murakami, H. (1994) The 14-3-
3 protein as the antigen for lung cancer-associated human monoclonal antibody AE6F4.
Hum. Antibod. Hybridomas 5 , 123-130.

Discussion 623
Barteling:
Do you have any idea what makes the individual background of the
Ichikawa: cells that you use? What gives the difference in dependency on
IL-2 and IL-4? Is it receptors, or efficiency of uptake, maybe?
We think that the ratio of B and T cells and other accessory cells in
our in vitro immunisation system will affect immunisation
efficiency and cytokine sensitivity of each cell population which are
maybe different depending upon donors. Determining the optimum
concentration of R2 and RS4 is useful to overcome the problem of
donor difference.

SESSION ON :
NEW TECHNOLOGIES FOR HEALTH CARE PRODUCTS
The use of ex vivo grown animal cells for cell therapy or for the establishment of bio-
artificial organs needs coculture techniques and a deep view into the metabolic
interactions between cells, cell and matrices and cells within a tissue. The different
aspects are covered by studying the extension of human hematopoietic progenitor cells in
coculture with stroma cells or by studying the metabolic competence of primary porcine
hepatocytes in the presence of endothelial cells and Kupfer cells. The animal imaging and
spectroscopy studies in hollow fibre bioreactros offers an excellent possibility to measure
the interactions of cells under tissue-like densities. The capability of cells to migrate
within the tissue is difficult to measure. Therefore, new methods for its determination are
of high interest.
J. Lehmann, E. T. Papoutsakis
Chairpersons

625

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 625.
1998 Kluwer Academic Publishers. Printed in the Netherlands.

ANALYSIS OF CELL GROWTH IN A FIXED BED BIOREACTOR
USING MAGNETIC RESONANCE SPECTROSCOPY AND IMAGING

PETER E. THELWALL, MARIA L. ANTHONY, DIETER
KEVIN M. BRINDLE

University of Cambridge
Department of Biochemistry, 80 Tennis Court Road, Cambridge
CB2 1GA, U.K.
‡Techinical University of Hamburg-Harburg
Bioprocess and Biochemical Engineering
Denicke Str. 15, 21071 Hamburg, Germany

1. Introduction

Intensive bioreactors, in particular the hollow-fibre bioreactors (HFBR), are widely used

for the production of proteins from mammalian cells [1-3] and for the growth of human
and animal cells ex vivo for cellular therapies, biohybrid artificial organs and tissue
engineering applications [4-6]. The principle problem with intensive reactor systems is
the formation of nutrient and waste product gradients [1]. These could have a negative
effect on cell growth and also, potentially, the quality of a protein product in terms of
its homogeneity. For example culture pH has been shown to affect the glycosylation
patterns of a secreted protein [7]. Perturbation of cellular metabolism by these gradients
could be particularly important in bioartificial organs, such as the bioartificial liver,
where the metabolic activity of the tissue is intimately related to its function.

Improvements in the design and operation of these systems require techniques
for monitoring the levels of nutrients and waste products and the cell distribution in
specific regions of the reactor. The closed nature of the reactors make them intrinsically
inaccessible to conventional monitoring techniques. For example studies of oxygen
distribution, which is frequently a limiting nutrient in these systems, have relied on
indirect methods or invasive electrode measurements [8-10], which have the potential to
disrupt reactor function. There is a need, therefore, to develop non-invasive techniques
for monitoring reactor performance. Magnetic resonance imaging (MRI) and
spectroscopy (MRS) are non-invasive techniques which have already been used to
monitor cellular metabolism [11,12], cell density [13] and medium flow rates [14]
within HFBRs. We have shown that diffusion-weighted imaging techniques can be
used to map cell distribution in an HFBR [15] and that MR relaxation time
imaging measurements, on a perfluorocarbon probe molecule, can be used to map
oxygen distribution [16]. We show here that diffusion-weighted imaging and
spectroscopy can be used to monitor cell distribution and cell volume respectively in a
fixed bed bioreactor composed of macroporous carriers.

627

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 627-633.
1998 Kluwer Academic Publishers. Printed in the Netherlands.

628

2. Materials and Methods

2.1. MATERIALS

Enzymes, cell culture media and foetal calf serum were obtained from Sigma Chemical
Company (Poole, Dorset, U.K.).

2.2. CELL CULTURE

Chinese hamster ovary cells (CHO K1) were cultured in DMEMcontaining
glucose supplemented with 10% foetal calf serum, 2 mM glutamine, 100 units
penicillin and 100 streptomycin. The bioreactor system consisted of a stirred
tank fermenter (FT Applikon, Tewkesbury, U.K.), containing 1.25 l of medium, which

was pumped via water-jacketed tubing at to the bioreactor positioned in the NMR
magnet. The bioreactor consisted of a 25 ml volume fixed bed, composed of
carriers (Biomaterials Co., Japan). To innoculate the bioreactor,
cells were harvested from dishes and suspended in 200 ml medium. This was
pumped through the reactor until the majority of the cells had adhered. The seeding
density was thus approximately The reactor was then perfused

with medium from the fermenter at a flow rate of 25 ml Medium in the
fermenter was replaced by a continous bleed and feed system, with a maximum rate of
300 ml per day. The medium was gassed with by passing these gases

through silicone rubber tubing wound on a former surrounding the impeller blade of the
fermenter. The of the medium was monitored using a polarographic oxygen

electrode (Mettler Toledo, Leicester, U.K.) and the pH using a glass electrode (Broadley
James Co., Santa Ana, CA.). Oxygen tension and pH were controlled by regulating the
composition of the gas stream. The pH was maintained at 7.35 and the oxygen tension

at 100%. The temperature was maintained at 37°C.

2.3 NMR METHODS

Experiments were performed using a Varian Unity Plus 400 MHz wide-bore
spectrometer equipped with an unshielded gradient set. spectra and images were
acquired using a Varian 25 mm 1H imaging probe and spectra using a Bruker
25 mm probe. Conventional spin echo images (TE = 20 ms; TR = 3 s) were
acquired from transverse slices of the bioreactor using 256 phase encode increments.
The in-plane resolution was 0.1 x 0.1 mm and the slice thickness 2 mm. Diffusion-
weighted images were acquired using a stimulated echo (STEAM) sequence [15], which

included a pair of pulsed magnetic field gradients of 0.2 T and 2.5 ms duration in a
TE period of 40 ms. The mixing time (TM) was 0.3 s. Sixty four phase encode
increments, with 128 transients per increment, were acquired with a repetition time of
2 s. The slice thickness was 2 mm and the in-plane resolution was 0.1 x 0.4 mm.
Diffusion-weighted spectroscopy was performed using a STEAM sequence
with a pair of pulsed magnetic field gradients of 2.5 ms duration in a TE period of 40
ms. The mixing time was 0.3 s. Fifteen gradient strengths between 0.02 and 0.3 T
were used. The intracellular and extracellular water fractions were calculated as
described in [17].

629
NMR spectra were acquired using a 40 µsec, 90° pulse and a pulse
repetition time of 1.4 s. The spectra, which were the sum of 2000 transients, were
acquired into 8000 data points, with a spectral width of 10 kHz. Chemical shifts and
signal intensities were referenced to the resonance of methylene diphosphonate (MDP),
which was contained in a capillary tube within the fixed bed reactor.

3. Results

Figure 1 shows a MR spectrum from the fixed bed bioreactor, 14 days after cell
seeding. The chemical shift or frequency of the inorganic phosphate peak, which
was predominantly due to extracellular phosphate, can be used to estimate extracellular
pH within the reactor [12]. This indicated that the reactor pH was 7.04, which was
signficantly lower than the pH in the feed vessel, which was determined to be 7.35,
using a glass electrode.

Figure 2 shows the increase, with time, of the volume of a water fraction in
the reactor which has a low apparent diffusion coefficient (ADC). This fraction with
low ADC has been assigned previously to intracellular water [15,17]. The volume of
this fraction was determined using a diffusion-weighted MR spectroscopy experiment
[17] on a region of the reactor which was similar to that interrogated in the MR
experiment. The increase in cell volume, as determined from this increase in
intracellular water fraction, showed a good correlation with the MR measurements
of ATP content (fig. 3). On day 10 of the culture the media exchange rate was increased
from 150 mL to 300 mL . This resulted in an increase in cell growth which
was apparent in the ATP measurements. The figure quoted for the percentage of

630
intracellular water in the reactor is only approximate since its calculation depends on
knowing the and relaxation times of intra- and extracellular water. These have
not been determined and have been assumed to be equal for intra- and extracellular water.

Figure 4 shows a conventional spin echo image and a diffusion-weighted
image of water protons from the same transverse slice through the reactor. The
position of the carriers can be clearly identified in the conventional image. The
diffusion-weighted image shows water with a low ADC, i.e. intracellular water [15].
High signal intensity in this image corresponds to high cell density. The image shows
that cell growth in the reactor is very heterogeneous and maximal at the periphery of the
carriers.

631

4. Discussion

is a relatively new macroporous cellulose carrier which can support the
growth of adherent and non-adherent cell lines. The surface has been modified with
polyethyleneimine (PEI), which imparts a slight positive charge that enhances the
attachment of non-adherent cells, such as hybridoma [18]. The pores are larger than
100 µM and this enhances mass transfer of oxygen and other nutrients into the carrier.
The large size of the carriers used in this study ( 5 x 5 x 5 mm) allowed packing of the
carriers without the aggregation and channelling of medium flow that we have
experienced with other, smaller, carriers.

The diffusion-weighted images (figure 4) showed that cells grew to a depth of
~ 1 m m from the surface of the carriers and that the cell density was increased around the
cavities between the carriers. These cavities are assumed to contain regions of increased
medium flow. Although not determined in this study, this can easily be investigated
using flow-sensitive imaging methods [14].

An important parameter to determine in metabolic studies of intensive
bioreactor systems is the viable cell concentration. A non-invasive technique is required
as sampling the reactor would be difficult and involve considerable disruption of the
culture. Previous studies have made estimates based on nutrient uptake rates [1].
However this assumes constant specific uptake rates, which may not be maintained
when conditions in the reactor are changed (discussed in [13]). An NMR-based method
has been described in which MR was used to measure total sodium ion
concentration in an HFBR [13]. As the cells grew and filled the extracapillary space of
the reactor, the sodium NMR signal declined as theintracellular concentration
(~ 7 mM) was very much less than the extracellular concentration (158 mM). This
increase in cell volume in the reactor showed a good correlation with the ATP content
determined by MR. However the problem with this experiment is that it involves
measuring a relatively small decrease in intensity of the sodium signal and will not be
very good, therefore, at detecting cells present in low numbers. The cell volume
determination experiment described here, in which the intracellular volume in the reactor
was determined by measuring the fraction of water with low apparent diffusion
coefficient (ADC) is, in principle, more sensitive. The main problem with the
experiment is that there is a small fraction of water, which is not intracellular, but
nevertheless has a low ADC. However this fraction, which is presumably water trapped
within the carrier matrix, can readily be determined by making measurements on the
reactor prior to cell loading (see fig. 2). The increase in the apparent intracellular
volume in the reactor (fig. 2) shows a good correlation, under these reactor conditions,
with the increase in ATP concentration determined by MR measurements (compare
figs. 2 and 3).

MR spectroscopy and imaging are clearly useful tools for investigating
intensive bioreactor performance. Conversely, intensive reactors are invaluable to those
wishing to use MR spectroscopy to investigate the metabolism of cultured cells.
Cultured cells, derived from specific tissues e.g. brain, tumours etc. have been used by
MR spectroscopists to generate relatively homogeneous and well defined tissue models.
MR studies on these systems can then be used to increase our biochemical
understanding of the MR data generated from studies on human tissues in the clinic.
Although short term studies on immobilised cell systems have been used, intensive
reactor systems, such as the HFBR, have the distinct advantage that they allow MR

632

measurements to be made on actively growing cells and are therefore likely to be more
physiologically relevant. The fixed bed reactor described in this study is particularly
attractive as it is easy to construct and can be sterilised by autoclaving. Hollow-fibre
reactors, on the other hand, require specialist skills to fabricate, are expensive to
purchase in a configuration which is compatible with use in a high field NMR
instrument and may require more sophisticated sterilisation techniques.

5. Acknowledgements

The work was supported by the European Community Framework IV Programme
(Biotechnology 950207). PET thanks the MRC for a studentship. Thanks to
Biomaterials Co. for providing the carriers.

6. References

1. Chresand, T. J., Gillies, R. J., and Dale, B. E.: Optimum fiber spacing in a hollow fiber bioreactor,
Biotech. Bioeng. 32 (1988), 983-992.

2. Knazek, R. A., Gullino, P. M, Kohler, P. O., and Dedrick, R. L.: Cell culture on artificial capillaries:
An approach to tissue growth in vitro., Science 178 (1972), 65-67.

3. Knight, P.: Hollow fiber bioreactors for mammalian cell culture, Bio/Technology 7 (1989), 459-461.
4. Hubbell, J. A., and Langer, R.: Tissue engineering, Chem. Eng. News 73 (1995), 42-54.

5. Jauregui, H. O., Chowdhury, N. R., and Chowdhury, J. R.: Use of mammalian liver cells for artificial
liver support. Cell Transplantation 5 (1996), 353-367.

6. Langer, R., and Vacanti, J. P.: Tissue engineering, Science 260 (1993), 920-926.
7. Borys, M. C., Linzer, D. I. H., and Papoutsakis, E. T.: Culture pH affects expression rates and

glycosylation of recombinant mouse placental lactogen proteins by Chinese Hamster ovary (CHO)
cells, Bio/Technology 11 (1993), 720-724.
8. Drury, D. D., Dale, B. E., and Gillies, R. J.: Oxygen transfer properties of a bioreactor for use within
a nuclear magnetic resonance spectrometer, Biotech. Bioeng. 32 (1988), 966-974.
9. Piret, J. M., and Cooney, C. L.: Model of oxygen transport limitations in hollow fiber bioreactors,
Biotech. Bioeng. 37 (1990), 80-92.
10. Wiesmann, R., Maier, S. T., Marx, U., and Buchholz, R.: Characterization of oxygen transfer in a
membrane-aerated hollow-fibre bioreactor using modified microcoaxial needle electrodes, Appl.
Microbiol. Biotechnol. 41 (1994), 531-536.
11. Gillies, R. J., MacKenzie, N. E., and Dale, B. E.: Analyses of bioreactor performance by nuclear
magnetic resonance spectroscopy, Bio/Technology 7 (1989), 50-54.
12. Gillies, R. J., Scherer, P. G., Raghunand, N., Okerlund, L. S., Martinez-Zaguilan, R., Hesterberg, L.,
and Dale, B. E.: Iteration of hybridoma cell growth and productivity in hollow fiber bioreactors using
31P NMR, Magn. Reson. Med. 18 (1991), 181-192.
13. Mancuso, A., Fernandez, E. J., Blanch, H. W., and Clark, D. S.: A nuclear magnetic resonance
technique for determining hybridoma cell concentration in hollow fiber bioreactors, Bio/Technology 8
(1990), 1282-1285.
14. Hammer, B. E., Heath, C. A., Mirer, S. D., and Belfort, G.: Quantitative flow measurements in
bioreactors by nuclear magnetic resonance imaging, Bio/Technology 8 (1990), 327-330.
15. Callies, R., Jackson, M. E., and Brindle, K. M.: Measurements of the growth and distribution of
mammalian cells in a hollow-fiber bioreactor using nuclear magnetic resonance imaging,
Bio/Technology 12 (1994), 75-78.
16. Williams, S. N. O., Rainer, R. M., and Brindle, K. M.: Mapping of oxygen tension and cell distribution
in a hollow fiber bioreactor using magnetic resonance imaging, Biotech. Bioeng. 56 (1997), 56-61.
17. van Zijl, P. C. M., Moonen, C. T. W., Faustino, P., Pekar, J., Kaplan, O., and Cohen, J. S.: Complete
separation of intracellular and extracellular information in NMR spectra of perfused cells by
diffusion-weighted spectroscopy, Proc. Natl. Acad. Sci. U.S.A. 88 (1991), 3228-3232.
18. Ong, C. P., Portner, R., Markl, H., Yamazaki, Y., Yasuda, K., and Matsumura, M.: High density
cultivation of hybridoma in charged porous carriers, J. Biotechnol. 34 (1994), 259-268.

Discussion 633
Noé:
Brindle: What is the interpretation of the peaks on your last slide?
The interesting ones are from phosphomonoester compounds which
Al-Rubeai: are membrane precursors which change in various disease states, and

Brindle: also when the cells are proliferating there are characteristic increases
in their levels.
Al-Rubeai: It is interesting to see the relationship between apoptosis and
acidification, and that when cells are subjected to hypoxia they
Brindel: undergo acidification.

They do that anyway. I would not say it is due to apoptosis. With
agents such a Cerimide, which induce apoptosis, you do see
acidification.

Could not this be due to changing the cell cycle as cells in G1 are
more acid than cells in S and G2 where you have alkalisation before
cell division?
We have tried to synchronise cells but with little success. The
interesting part was that a fraction of the cells apparently died and
the rest survived in anoxic conditions.

EXPANSION OF HUMAN HEMATOPOIETIC PROGENITOR CELLS IN A
FIXED BED BIOREACTOR

P. MEISSNER, P.WERNER°, B. SCHRÖDER*, C. HERFURTH,
C. WANDREY, M. BISELLI
Institute of Biotechnology, Forschungszentrum Jülich, °Stem Cell Bank,
Heinrich-Heine-Universität, *ASTA Medica, Frankfurt, Germany.

Abstract: The ex vivo expansion of hematopoietic progenitor cells is very important
for a variety of clinical applications, e.g. bone marrow transplantation or gene therapy
Therefore it is of general interest to develop a culture system able to mimic the in vivo
hematopoietic environment.
We have modified a continuously perfused bioreactor originally developed for the
production of monoclonal antibodies. In a fixed bed reactor we immobilized stromal
cells (human primary stromal cells or the murine stromal cell line M2-10B4) in porous
glass carriers and inoculated human hematopoietic progenitor cells two days later and
cocultivated them for several weeks. At different times after inoculation of
mononuclear cells (MNCs) derived from umbilical cord blood or peripheral blood stem
cells both adherent and non adherent cells were harvested and analyzed by
flowcytometry, short-term colony assays, and LTC-IC analysis.
During the cultivation there was a permanent production of progenitor cells and mature
blood cells derived from the immobilized cells in the carriers. We could demonstrate the
expansion of CFU-GM (7-fold), BFU-E (1.8-fold), and CFU-GEMM (4.2-fold) in

contrast to the maintenance of late progenitors (CFU-E). Additionally we could
observe the immobilization and 4.8-fold expansion of LTC-ICs.

Discussion

Hatzfeld: When you say you have expansion of CFU-GM cells did you really
check that the CFU-GM cells at day 1 provide the same size colony
as the CFU-GM you observe at day 8?

Meissner: The distribution of the colony size was the same on day 0,1 and day
8, and therefore we have a real expansion

Hatzfeld: Are you sure the cells were CFU-GM and not CFU-E? Have you
checked for granulocytes, etc?

Meissner: Our colonies have no CFU-E cells at the periphery. There are white
cells, monocytes and granulocytes at the periphery

635

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 635-636.
1998 Kluwer Academic Publishers. Printed in the Netherlands.

636 Regarding the starling material - you mentioned immobilised cells -
Ostrove: what was the actual starting material in the bioreactor and were they
CD34 selected?

Meissner: We started only with mononuclear cells.

Ostrove: Do you have any evidence that CD34 will grow in the reactor?

Meissner: We can show immobilisation of CD34 positive cells but there is no
selection as cells are in the medium as well.

Scheirer: How do your results compare with standard monolayer cultures in
bottles with feeder layers?

Meissner: You cannot compare these expansion results with normal co-cultures
because it is a perfusion system so the media change rate is not the
same. Also we did not add cytokines.

Anon: Did you also look at the more primitive CD34 and CD38 cells?

Meissner: No.

Lowagie: Did you try other immobilisation methods - microcarriers for
example?

Meissner: No.

Lehmann: Did you check the medium requirements for different cells and then
design the medium, or was it just by chance?

Meissner: We used the typical Dexter type medium only to see if the reactor
system was OK. This was because we used this medium before for

co-cultivation of the stroma cells with CD34 positive hemopoetic
cells, so we did not change the medium or add any cytokines.

Scharfenberg: How homogeneous is your fixed bed - is it homogeneous the whole
time, or for a number of weeks?

Meissner: Yes, it is homogeneous because the circulation pump gives a
permanent medium flow.

Scharfenberg: To clarify the question - is it homogeneous with respect to the
different cell types?

Meissner: When we take samples it is only from the bottom of the bead bed, so

I cannot comment on your question.

A NOVEL ASSAY TO DETERMINE AND QUANTIFY THE
REGULATION OF CELL MOTILITY AND MIGRATION
DEMONSTRATED ON HEMATOPOIETIC CELLS

D.Möbest1),2), S. Ries1), R.Mertelsmann1), R. Henschler1)

1) Dept. of Hematology/Oncology, University Medical Center, Freiburg, FRG
2) Dept. of Biology, University of Freiburg, Freiburg, FRG

Hematopoiesis occurs in the bone marrow and is supported by a complex network of

extracellular matrix (ECM) and stromal cells (1). Hematopoietic cells including

developing progenitors have been shown to specificially interact with matrix proteins and

adhesion molecules on fibroblasts and other stromal cells. Progenitor-stromal cell

interactions may occur through the vascular cell adhesion molecule (V-CAM) expressed

on fibroblasts with the integrin receptor on hematopoietic cells, or through the

interstitial (I)-CAM 1 molecule on macrophages with integrins (2,3). The extracellular

matrix protein fibronectin (FN), through various domains, can bind to

integrins expressed on hematopoietic progenitor cells, whereas laminin (Lam) may bind

progenitors through the receptor (2,4). In long-term bone marrow cultures

(LTBMC), very primitive progenitor cells selectively bind to fibronectin and thus avidly

adhere to the stromal matrix (5). Moreover, monoclonal antibodies against the

integrin, when applied to non-human primates, induce hematopoietic progenitor cells to

leave the bone marrow and circulate in blood (6). Blocking antibodies against the CD44

(Hermes) antigen resulted in decreased lympho-hematopoiesis within LTBMC (7).

Levesque et al. (8) showed that cytokines rapidly upregulate the avidity of the

integrins on hematopoietic progenitors, resulting in their adhesion to FN or Lam. Since
during the homing of the hematopoietic stem and progenitor cells, e.g. after bone marrow
transplantation in patients, the recolonizing cells have to traverse the endothelial barrier

637

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 637-643.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

638

and subsequently the stromal interstitial space within the bone marrow, cell migration on
extracellular matrix proteins is very likely to be intricately involved in the reinitiation of
blood cell formation after stem cell transplantation (2,3). We therefore investigated
possibilities to quantify and further characterize the behaviour, especially the migration
and motility, of hematopoietic progenitors on ECM by studying migration on tilted,
matrix-coated cell culture plates.

Resistance or shear forces, which occur in vivo and which have to be overcome by
circulating blood cells, have been previously introduced into in vitro culture systems:
Transwell pores require active movements of cells through artificial lumina; flow
chambers allow to study the initial adhesion steps e.g. of selectin binding (9). We
analyzed migration on ECM coated plates using photography or video microscopy. For
the assays, FDCP-mix cells (Factor dependent cells Paterson-mixed potential (10)) were
seeded at 10.000 per well on FN or Lam- precoated (10 µg/cm2) 96 well plates in IMDM

and 20% horse serum. As shown in Fig.1, cells were allowed to accumulate at one end of
the well by tilting the plates at an 80° angle for 24 h.

639

Subsequently, at timepoint 0 h of experiments, cells were allowed to move after lowering
the tilt angle to 15°. It was noticed that the progenitors spread uphill from about 6 h after
start of the experiment, reaching plateau levels of cell migration after 15-24h (Fig.1).
Video microscopy revealed that cell movements occured at random directions. As a first
endpoint, the number of cells that had moved away from the upper edge of the coherent
cell mass that located at the bottom of the wells was quantified. As a second endpoint, the
velocity of individual cells was determined.

A variety of cytokines and chemokines was investigated showing stimulatory, inhibitory

and indifferent effects (Table 1). A dose response of hematopoietic cytokines granulocyte

macrophage-colony stimulating factor (GM-CSF) and Interleukin (IL)-3 could be shown.

Variation of ECM concentration and inclination angle on FN, Lam resulted in dose-

dependent effects, respectively. An optimal resolution of these effects could be seen at a

coating concentration of 10 and an inclination of (Table 1). Anti-integrin

antibodies against the and the chain abrogated migration on FN, and anti or
anti- on Lam (Tab.l). The activation time of migration was dependent on exposure

time to gravity force, but not on the exposure time of IL-3. Migration speed depended on
FN, but not on IL-3 concentration. A sudden incrase of the IL-3 concentration resulted in

an sharp decrease of the motility of cells, which was reversible over time. The net

potential energy, which could be achieved by migrating cells was higher on FN coated

surfaces as compared to BSA coated surfaces.

We observed that our migration assay was able to detect and quantify the migrational
behaviour of hematopoietic progenitor cells, which was mainly influenced by cytokines
and ECM. In particular, this assay induces a restraint force onto the cells by simple tilting
of the plates. One main result was that the kinetics determined in this migration assay are
different from the induction of adhesion or induction of migration on even surfaces (11).
The delay observed may be due to the requirement of intracellular reorganization of the
cytoskeleton for example, posttranslational modifications which may have to take place
before a coordinated migration against gravity force is inducable. In our assay the
"antigravity" migration was dependent on integrins, whereas migration on even
surfaces was not (11). Interestingly, a restraint force on integrins has been induced in
fibroblasts by withholding FN coated beads that bound to integrins with an optical trap

(12,13). Possibly, in both situations an extracellular force is required to induce a
coordinated activation of focal adhesion complexes, or related structures, which are
neccessary to induce migrational movements.

640

641

This assay has been used to make visible the transition of a cellular integrin-mediated
migrational into an adhesional state (Tab.1). To our knowledge, this has so far not been
possible with conventional assays. This method will therefore be of advantage to possibly
dissect the role of certain G-protein families, such as rho, rac, or cdc42, GTP-ase in these
phenomena (14). An advantage of this assay is that it uses the same culture equipment
that can be applied in classical adhesion assays, thus allowing additional comparison to
established technology. Compared to the flow chamber assay, our assay is technically
relatively more simple. To address the question how progenitor cells find their
appropriate niche within the bone marrow stroma, this assay may be more suitable,
whereas the flow chamber assay has originally been designed to imitate the situation in
blood stream. Therefore, this system allows screening experiments, especially if it is
combined with automated cell scoring systems. In addition it is a very open system,
allowing quick qualitative answers, but can also be used to achieve quantitative results
e.g. regarding motility of cells, energy of cells used for migration, or the effect of
external force on the dynamics of cells.

References

1. Dexter, T.M., Allen, T.D., and L.G. Lajtha.: Conditions controlling the proliferation

of hematopoietic cells in vitro. J. Cell. Physiol. 91:335-344, 1977
2. Hynes, R.O.: Integrins: versatility, modulation, and signalling in cell adhesion. Cell

69:11-25, 1992
3. Springer, T.M.: Traffic signals on endothelium for lymphocyte recirculation and

leukocyte emigration. Annu. Rev. Physiol. 57:827-872, 1995
4. Williams, D.A., Rios, M., Stephens C., Patel V.P.: Fibronectin and VLA-4 in

haematopoietic stem cell-microenvironment interactions. Nature 352:438-441, 1991
5. Verfaillie, C., Blakolmer, K., McGlave, P.: Purified primitive human hematopoietic

progenitor cells with long term repopulating capacity adhere selectively to irradiated bone
marrow stroma. J. Exp. Med. 172:509-520, 1990
6. Papayannopoulou, T., Nakamoto, B.: Peripheralization of hemopoietic progenitors in
primates treated with anti-VLA4 integrin. Proc. Natl. Acad. Sci. U.S.A. 90:9374-9378,
1993
7. Miyake, K., Medina K.L., Hayashi S.I., Ono, S., Hamaoka, T., Kincade, P.W.:
Monoclonal antibodies to Pgp-1/CD44 block lympho-hematopoiesis in long-term bone
marrow cultures. J. Exp. Med. 171:477-488, 1990

642

8. Levesque, J.-P., Leavesley, D.I., Niutta, S., Vadas, M., Simmonds, P.J.: Cytokines
increase human hematopoietic cell adhesiveness by activation of very late antigen (VLA)-4
and VLA-5 integrins. J. Exp. Med. 181:1805-1815, 1995

9. Chen, S., Alon, R., Fuhlbrigge, R.C., Springer, T.A.: Rolling and transient tethering of
leucocytes on antibodies reveal specializations of selectins. Proc. Natl. Acad. Sci. U.S.A.
94(7):3172-3177, 1997

10. Spooncer, E., Boettiger, D., Dexter, T.M.: Isolation and culture of factor-dependent
haemopoietic cell lines, in N.G. Testa, and G. Molineux (eds): Haemopoiesis - a practical
approach. Oxford (UK), Oxford University Press, 1992, p 106-121

11. Strobel, E.-S., Möbest, D., von Kleist, S., Dangel, M., Mertelsmann, R., Henschler,
R.: Adhesion and migration are differentially regulated in hematopoietic progenitor cells
by cytokines and extracellular matrix. Blood 90:3524-3532, 1997.

12. Felsenfeld, D.P., Choquet, D., Sheetz, M. P.: Ligand binding regulates the directed
movement of integrins on fibroblasts. Nature 393:438-440, 1996

13. Choquet, D., Felsenfeld, D. P., Sheetz, M. P.: Extracellular matrix rigidity causes

strengthening of integrin-cytoskeleton linkages. Cell 88:39-48, 1997
14. Nobes, C.D., Hall A.: Rho, rac, and cdc42 GTPases regulate the assembly of

multimolecular focal complexes associated with actin stress fibres, lamellipodia, and
filopodia. Cell 81:53-62, 1995

Discussion 643
Papoutsakis:
What was quantitatively different between the cell line and the CD34
Möbest: positive cells which you saw at the end?
So far we have not found any differences but all the results have not
Papoutsakis: yet been analysed.
Möbest: Would you expect to see any differences?
The CD34 cells look very similar but there may be some differences
in the sub-populations.

IMMORTALIZATION OF DIFFERENTIATED HEPATOCYTES

GARY S. JENNINGS* AND MICHAEL STRAUSS§
*HepaVec AG and §Humboldt University of Berlin
Robert-Roessle-Strasse 10, D-13122 Berlin, Germany

1. Introduction

1.1. IMMORTALIZED HEPATOCYTES

The potential uses for continuously proliferating hepatocyte-derived cells have mutiplied
in the last thirty years but with this the demand for ever more differentiated functions
has grown louder. Amongst the most vociferous lobbies are those of xenobiotic
toxicology and recombinant protein technology, two fields where human liver-specific
metabolism and polypeptide processing would be advantageous. The development of a
bioartificial liver using hepatocyte cell lines instead of primary cells is another
application which especially interests us. While current ex vivo liver-support devices
containing porcine hepatocytes have been shown to maintain some degree of the
complexity of liver functions [1], the low proliferative capacity and rapid loss of specific
enzyme activities in primary culture limits their use to the very short-term [2].
Moreover, in the longer term the use of non-human hepatocytes may not confer the
required specificity of metabolism.

In seeking a cell line suitable for clinical application we decided to avoid those
stemming from hepatoma material, several of which produce infectious hepatitis B virus
[3-7], and others derived from cells expressing immortalizing viral antigens, such as the
large T-antigen of SV40 [8,9] or the E1 antigens of human adenoviruses (Table 1). In
both cases, despite the maintenance of many liver-specific activities, levels of most are
decimated [2]. Furthermore, foetal antigens such as -fetoprotein may be reinduced
indicating a dedifferentiated phenotype and indeed, many of these cell lines are
tumorigenic [9]. However, analysis of changes in two tumour suppressor genes involved
in cell division gives us a clue as to how we might stimulate proliferation in normal
human hepatocytes in a controlled fashion.

The product of the retinoblastoma susceptibility gene (RB1) is a critical regulator of
cell cycle entry and the protein is involved in gating cells for apoptosis. In the
majority of hepatoma cell lines and in all cell lines expressing immortalizing viral
antigens one or both of the tumour suppressors is deregulated (Table 1).

645

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 645-655.
© 1998 Khiwer Academic Publishers. Printed in the Netherlands.

646

1.2. THE RETINOBLASTOMA PROTEIN IN CELL CYCLE ENTRY
E2F-1 transcription factor is synthesized in the mid- phase of the cell cycle and is
required for induction of genes in late- and S phase. Levels of free E2F-1 are
regulated through sequestration by the retinoblastoma protein [12,13], which thereby
prevents transcription from E2F-1 binding promoters and also enters into an active
suppressor complex with specificity for a different class of promoters [14,15]. In this
way pRb coordinates induction of genes during to S transition (Fig. 1). The pRb
checkpoint is also the site at which positive and negative growth signals are integrated
and the decision taken to cross the restriction (R) point [14,15]. Mitogenic signalling
results in the induction of D-type cyclins [16], the regulatory subunits of cyclin-
dependent kinase 4 (cdk4). The active kinase hyperphosphorylates and inactivates pRb,
causing release of free E2F-1 (Fig. 1). Another class of proteins, the cyclin-dependent
kinase inhibitors (CKI), can bind to and block the kinase activity of the cdks and one

647

such inhibitor, has cdk4 as its sole target [17]. Expression of appears to

be repressed by the pRb-E2F-l complex and this is relieved by hyperphosphorylation of

pRb [18]. In normal cells, cyclin Dl becomes degraded in leaving an inactive

complex which survives until cyclin D1 is resynthesized in the next
phase (Fig. 1).

In most tumour cell lines the pRb checkpoint is deregulated, either by functional

inactivation of pRb or , or by overexpression of cyclin D1 or, occasionally, cdk4

[14,19]. This enables cells to enter S phase with reduced growth factor requirement and
importantly removes a checkpoint essential for normal growth control. While
permanent changes in the pRb checkpoint may signal the initiation of neoplasia it is also
possible to generate hyperplasia in the liver. Stimuli which act to compromise liver
function, such as partial hepatectomy, induce the remaining cells to enter cell division.

New cells differentiate and assume the role of fully functioning hepatocytes but retain

the ability to proliferate when necessary [20]. By transiently modulating expression of
cell cycle regulators in vitro, we attempted to recreate hepatocyte hyperplasia and enable
the outgrowth of hepatocyte cell lines with intact and S phase checkpoints.

2. Transient neutralization of the retinoblastoma checkpoint

Previous work has shown that a transient reduction in the level of pRb in human
embryonic lung (HEL) fibroblasts leads to stimulation of proliferation in vitro (Fig. 2)

[21]. This was accomplished by transfection of phosphorothioate-based antisense

oligodeoxyribonucleotides, eighteen bases in length, with a sequence directed against

648
the human RB1 message. However, the high concentration of oligonucleotide required

and lower level of uptake by cultured human hepatocytes forced us to consider
an altenative strategy. We constructed a plasmid with a 500-nucleotide antisense
sequence from the human RB1 gene which covered the initiation codon, under control of
an albumin promoter/enhancer element (Fig. 3) which should be expressed when

transfected into adult hepatocytes. Additionally, plasmids containing cDNA sequences
for the human cyclin Dl and E2F-1 genes driven by modified viral promoters were
made (Fig. 3) and these were transfected singly or in combination with the antisense-
RB1 plasmid. Cultures of human hepatocytes transfected with the described plasmids
were maintained without selection in a serum containing medium. Non-transfected cells
ceased to divide after one week and at 6 weeks most cells were dead and had detached
from the plate. Over the following 2 weeks the outgrowth of epitheloid colonies was
observed (Fig. 4a) the number of these being greater in cotransfection experiments than
in single plasmid transfections. However, without exception, the colonies underwent a
wave of cell death at 8 to 12 weeks after transfection. In each case cell-cell contact was
lost and cell size decreased. Subsequently, nuclear condensation and fragmentation and
later plamsa membrane blebbing was evident (Fig. 4b), all of which are consistent with
an apoptotic phenotype.
3. Inhibition of p53-dependent apoptosis
Apoptosis is one consequence of contradictory growth signals in cells unable to arrest in

[22]. As discussed above, the majority of hepatoma cell lines and cells harbouring
immortalizing viral antigens also target the p53 protein for inactivation. Other viral

antigens, the E1B 19 kDa protein for example, are homologues of cellular anti-apoptotic

649
proteins [23]. It appears then that removal of the pRb-checkpoint necessitates
suppression of apoptosis to secure continuous proliferation. To this end we constructed a
fourth plasmid in which a 500-nucleotide antisense sequence from the human p53 gene
was placed downstream of the albumin promoter/enhancer element (Fig. 3). This

plasmid was cotransfected with one or more of the previous three plasmids into human

hepatocytes as before. Similar numbers of colonies were obtained as with earlier
transfections however, several colonies continued to expand after the majority had died
during the apoptotic stage. One of these colonies could be stably passaged and was
established as the cell line, HepZ (Fig. 4c).

4. Characterization of the immortalized hepatocyte line, HepZ
4.1. CELL CYCLE ENTRY
The HepZ cell line has been maintained for over 40 passages (approximately 120
population doublings) with no observable changes. Cells display a polygonal shape and
intercellular spaces reminiscent of canaliculi in primary hepatocyte culture (Fig. 4c). At
confluence, cell-cell contacts are remarkably strong whilst adherence to plastic dishes is
very weak and HepZ was routinely plated on collagen I-coated dishes. Survival in
culture depends on the presence of serum and growth rate is proportional to serum
content between 0.5-10% FBS. This indicates that the growth factor sensitive checkpoint
regulated by pRb is intact. Western blots of protein extracts from logarithmic phase

cultures show that pRb is amply present and exists primarily in the hyperphosphorylated

form (Fig. 5a). Cyclin D1 is also expressed at similar levels as in other proliferating
hepatocyte cell lines (Fig. 5b). These data are consistent with the effects in of
mitogenic signalling and not of genomic integration of antisense-RB1 or cyclin

650

plasmids. E2F-1 protein, on the other hand, is present at highly elevated levels when
compared to other immortalized cells or with regenerating mouse liver (Fig. 5c) and this

may be due to expression from an exogenous integrated plasmid. The fourth protein

whose synthesis might have been influenced by our intervention was p53. We performed

immunoprecipitations with mouse monoclonal anti-p53 antibodies specific for wild-type

or mutant protein conformations followed by Western analysis using a rabbit polyclonal
antibody. No protein was recognized by the wild-type specific antibody (data not
shown) but a strong band was produced with antibody specific for the mutant form (Fig.
5d). Western blotting of other to S phase proteins, cdk4, cdk2, cyclin E, cyclin A,
cdc25A and PCNA, and of the revealed no
abnormalities in expression (data not shown).

4.2. HEPATOCYTE SPECIFIC ACTIVITIES

Because of our interest in the use of viral vectors in gene therapy we next examined the
efficacy and specificity of infection of HepZ by adenovirus and baculovirus. Adenovirus
efficiently infects hepatocytes [24] and Fig. 6a shows that HepZ is also efficiently
infected by adenovirus carrying a -galactosidase gene. Recently we reported specific
infection of hepatocytes and hepatocyte-derived cell lines by baculovirus [25]. Infection
of HepZ by baculovirus (Fig. 6b) though not as efficient as that by adenovirus is at least

651
as high as in primary hepatocyte cultures [25]. We conclude that HepZ cells maintain at
least some viral specific cell-surface proteins present in hepatocytes.

Hepatocytes respond to glucocorticoids by increasing gluconeogenesis. The switch in
metabolic emphasis away from protein and DNA synthesis results in a slower

proliferation rate in cultured cells [26]. HepZ responded to addition of 0.1 µM

dexamethasone to the medium with a reduction in growth to around 40% of controls
(Fig 7). In further immunohistochemical characterization HepZ was found to produce
albumin but not -fetoprotein and to express several isoforms of cytochrome P450 (data
not shown).

5. Summary
The majority of hepatocyte cell lines are derivatives of neoplasias and contain growth
deregulating genetic abnormalities. Whilst some liver specific activities are still

detectable, their low levels make it unlikely that these cell lines could substitute for
hepatocytes in bioartificial liver systems hence new immortalized differentiated
hepatocyte lines are needed. The liver is capable of mounting a rapid proliferative
response to a loss of cell mass, either through degenerative disease or surgical
intervention. Hepatocytes produced during this compensatory hyperplasia later
differentiate but maintain the capacity to enter cell division. We attempted to induce a
similar hyperplasia in cultured primary hepatocytes by repressing tumour suppressor
protein action in the transition from to S phase. Plasmids bearing antisense-RB1 and

antisense-p53 sequences as well as plasmids encoding E2F-1 and cyclin D1 were
transfected into primary human hepatocytes in an attempt to create a transient
environment where cell cycle entry would be ameliorated. In transfections omitting the
antisense-p53 plasmid, all colonies produced underwent rapid cell death after 8 weeks

652

but in those including antisense-p53 a small number continued to grow, indicating that
transient reduction of p53 allows primary hepatocytes an extended proliferative phase.
One colony was passaged and established as the cell line HepZ.

HepZ cells maintain a hepatocyte-Iike phenotype judged by several criteria and may be
useful in the development of a liver support device incorperating immortalized polymer
cells. In this respect it has already been demostrated that the cells are able to grow on
beads (Jürgen Lehmann, personal communication) a feature enabling their culture at
high cell density, a prerequisie for a bioartificial liver system.

Western blotting shows that HepZ expresses elevated amounts of E2F-1 protein and
has a mutant p53. The former probably allows the cells rapid transit through the
checkpoint controlled by pRb and the latter escape from programmed cell death. This is
resemblant of cell lines derived from hepatomas or transformed with immortalizing viral
antigens in which the same control points are usurped. In these experiments no selection
was applied with the result that the fastest growing cells were isolated and this is likely
the reason for the acquisition of a mutation in the p53 gene. In future experiments a
selection pressure to maintain the integrity of checkpoints will be applied.

6. References

1. Gerlach, J.C.: Development of a hybrid liver support system: a review, Int. J. Artif. Organs 19 (1996),
645-654.

2. Rogiers, V., and Vercruysse, A.: Rat hepatocyte cultures and co-cultures in biotransformation studies of
xenobiotics, Toxicology 82 (1993), 193-208.

3. McNab, G.M., Alexander, J.J., Lecatas, G., Bey, E.M., and Urbanowicz, J.M.: Hepatitis B surface
antigen produced by a human hepatoma cell line, Br. J. Cancer 34 (1976), 509-515.

4. Aden, D.P., Fogel, A., Plotkin, S., Damjanov, I., and Knowles, B.B.: Controlled synthesis of HBsAg in a
differentiated human liver carcinoma-derived cell line, Nature (London) 282 (1979), 615-616.

5. Nakabayashi, I I . , Taketa, K., Miyano, K., Yamane, T., and Sato, J.: Growth of human hepatoma cell
lines with differentiated functions in chemically defined medium, Cancer Res. 42 (1982), 3858-3863.

653

6. He, L., Isselbacher, K.J., Wands, J.R., Goodman, H.M., Shih, C., and Quaroni, A.: Establishment and
characterization of a new human hepatocellular carcinoma cell line, In vitro 20 (1984), 493-504.

7. Alexander, J.J.. In vitro studies of human hepatocellular carcinoma cell lines. Adv. Hepatitis Res. (1984),
190-195.

8. Pfeifer, A.M.A., Cole, K.E., Smoot, D.T., Weston, A., Groopman, J.D., Shields, P.G., Vignaud, J-M,
Juillerat, M., Lipsky, M.M., Trump, B.F., Lechner, J.F., and Harris, C.C. Simian virus 40 large tumor
antigen-immortalized normal human liver epithelial cells express hepatocyte characteristics and
metabolize chemical carcinogens, Proc. Natl. Acad. Sci. USA 90 (1993), 5123-5127.

9. Woodworth, C.D., Kreider, J.W., Mengel, L, Miller, T., Meng, Y.L., and Isom, H.C.: Tumorigenicity of
simian virus 40-hepatocyte cell lines: effect of in vitro and in vivo passage on expression of liver-specific
genes and oncogenes, Mol. Cell. Biol. 8 (1988), 4492-4501.

10. Puisieux, A., Galvin, K., Troalen, F., Bresac, B., Marcais, C., Galun, E., Ponchel, F., Yakicier, C., Ji, J.,
and Ozturk, M.: Retinoblastoma and p53 tumor suppressor genes in human hepatoma cell lines, FASEB
J. 7 (1993), 1407-1413.

11. Hsu, I.C., Tokiwa, T., Bennett, W., Metcalf, R.A., Welsh, J.A., Sun, T., and Harris, C.C.: p53 gene
mutation and integrated hepatitis B viral sequences in human liver cancer cell lines, Carcinogenesis 14

(1993), 987-992.
12. Qin, X.Q., Livingston, D.M., Ewen, M., Sellers, W.R., Arany, Z., and Kaelin, W.G. Jr.: The transcription

factor E2F-1 is a downstream target of RB action. Mol. Cell. Biol. 15 (1995), 742-755.
13. Sellers, W.R., and Kaelin, W.G.: pRB as a modulator of transcription, Biochim. Biophys. Acta. 1288

(1996), M1-5.
14. Weinberg, R.A.: The retinoblastoma protein and cell cycle control, Cell 81 (1995), 323-330.
15. Herwig, S. and Strauss, M.: The retinoblastoma protein: a master regulator of cell cycle, differentiation

and apoptosis, Eur. J. Biochem. 246 (1997), 581-601.
16. Lukas, J., Bartkova, J., and Bartek, J.: Convergence of mitogenic signalling cascades from diverse classes

of receptors at the cyclin D-cyclin-dependent kinase-pRb-controlled Gl checkpoint, Mol. Cell. Biol. 16
(1996), 6917-6925

17. Serrano, M., Hannon, G., and Beach, D.: A new regulatory motif in cell cycle control causing specific

inhibition of cyclin D/CDK4, Nature 366 (1993), 704-707.
18. Lukas, J., Parry, D., Aagaard, L., Mann, D.J., Bartkova, J., Strauss, M., Peters, G., and Bartek J.:

Retinoblastoma-protein-dependent cell-cycle inhibition by the tumor suppressor p16, Nature (London)
375 (1995), 503-506.
19. Strauss, M., Lukas, J., and Bartek, J.: Unrestricted cell cycling and cancer, Nature Med. 1 (1995), 1254-
1246.
20. Fausto, N., and Weber, E.M.: Liver regeneration, in I.M. Arias, J.L. Boyer, N. Fausto, W.B. Jakoby, D.A.
Schachter and D.A. Shafritz, (eds.), The Liver: Biology and Pathobiology, third edition, Raven Press
Ltd., New York, (1994), pp. 1059-1084.
21. Strauss, M., Hering, S., Lubbe, L., and Griffen, B.E.: Immortalization and transformation of human
fibroblasts by regulated expression of polyoma virus T antigens, Oncogene 5 (1990), 1223-1229.
22. Yonish-Rouach, E., Grunwald, D., Wilder, S., Kimchi, A., May, E., Lawrence, J.J., May, P., and Oren,
M.: p53-mediated cell death: relationship to cell cycle control, Mol. Cell. Biol. 13 (1993), 1415-1423.
23. Rao, L., Debbas, M., Sabbatini, P., Hockenbery, D., Korsmeyer, S., and White, E.: The adenovirus E1A
proteins induce apoptosis, which is inhibited by the E1B 19-kDa and Bcl-2 proteins, Proc. Natl. Acad.
Sci. USA 89 (1992), 7742-7746.
24. Cristiano, R.J., Smith, L.C., Kay, M.A., Brinkley, B.R., and Woo, S.L.: Hepatic gene therapy, efficient
gene delivery and expression in primary hepatocytes utilizing a conjugated adenovirus-DNA complex,
Proc. Natl. Acad. Sci. USA 90 (1993), 11548-11552.
25. Hofmann, C., Sandig, V., Jennings, G.S., Rudolph, M., Schlag, P., and Strauss, M.: Efficient gene
transfer into human hepatocytes by baculovirus vectors, Proc. Natl. Acad. Sci. USA 92 (1995), 10099-
10103.
26. Vintermyr, O.K., and Doskeland, S.O.: Characterization of the inhibitory effect of glucocorticoids on the
DNA replication of adult rat hepatocytes growing at various cell densities, J. Cell. Physiol. 138 (1989),
29-37.

654 By using anti-sense to inactivate tumour suppressor genes, like p53
Discussion or Rb, are you not making this potentially tumourigenic?
Singhvi:
Jennings: The growth rate increases but if you passage them long enough,
then the anti-sense RG is degraded and they go back to the same
Noé: growth rate. It is this transient effect that we are trying to use to
Jennings: promote hepatocyte growth. Within the hepatocyte outgrowth
from a primary culture we can then use selection methods like
MacDonald: collogen gel sandwich, etc, and hopefully develop a cell line. The
Jennings: gel sandwich technique has been shown to work well.
Grammalikos:
Jennings: Could you give more information on the status of the biochemical
characterisation of these cells and their potential in artificial livers?
Hauser:
Various P450’s have been identified with 3a4 being highly
expressed. They express glucourinyl transferase but at what levels
and specificity, I am unsure. The cells were selected only on the
basis of growth, so the next step will be to select them on
specificity.

How stable is the expression of the P450’s?

The cells have been passaged 50 times and there was no difference
between passage 20 and 40.

A question on semantics. What do you mean by a non-transformed
cell line?

We are thinking in terms of clinical aspects and thus we want to
avoid tumourigenic cell lines. Transformation can result in a
reduction of hepatocyte characteristics. Transformation by T-
antigen has all sorts of other effects which reduce hepatocyte
specific function but promote a general protein synthesis, which is
not wanted either. We require a proliferating, hepatocyte specific,
non-malignant cell line which is immortalised.

A question of how you transform these cells. You have a transient
application of your transforming genes and suddenly you have a
stable cell line. What is your understanding at the molecular level
above this phase and is it reproducible?

Jennings: 655
People may get upset by the fact that this cell line has an integrated
Singhvi: E2F-1 plasmid. You have to do something to the cells to make
Jennings: them grow. If we reduce the cell cycle blockers pRb53 low
enough for them to proliferate for a long time, or to use extra
special medium, somewhere along the line they will probably fix a
mutation and become immortalised. If this mutation is not
deleterious to our requirements, then it is OK for us. If the
mutation is in ras for instance and the whole cell cycle is intact,
then we would not be able to use this for clinical applications.
Ultimately, in a combination of very complicated selection factors,
we may be able to get the cells to proliferate without integrating
mutations. In regenerating mouse liver there are no mutations but
all the cells proliferate and then switch off when it gets to the right
size. We would like to do this in culture.

Your comment on liver coming to the right size, is this not
regulated on a cellular level? The size of the cells regulates
whether it is differentiated or growing.

If you transplant a mouse liver into a rat, the liver gets bigger but
the cells do not. If you transplant a rat liver to a mouse, the cells
remain the same size but the liver shrinks by apoptosis. This is a
systemic regulation and many factors control proliferation and
differentiation.

FIXED-BED REACTORS FOR ANIMAL CELL CULTIVATION:
AN APPROACH TO ARTIFICIAL ORGANS

R. PÖRTNER1, S. RÖSSING1, J. STANGE2, D. FASSNACHT1
1 Technische Universität Hamburg-Harburg, Bioprozeß- und
Bioverfahrenstechnik,DenickeStr. 15,D-21071 Hamburg, Germany
2 Universität Rostock, Klinik und Poliklinik für Innere Medizin,
Ernst-Heydemann-Str. 6, D-18055 Rostock, Germany

1. Abstract
In this study the immortalised mouse hepatocyte line mHep-R1 was used for

cultivation. The cells were first grown in culture flasks and in a small-scale fixed-bed

system in order to determine growth characteristics and a suitable carrier type. The cell
line was then cultivated in a 40 ml fixed-bed reactor over a period of 75 days at a
perfusion rate of 6.25 ml medium per ml fixed-bed and day. A cell density of approx.

cells per ml carrier was reached at the end of the experiment proving that a
stable cultivation was possible over a long period of time with constant consumption
and production rates.

2. Introduction
The immobilisation of animal cells in fixed-bed systems proved to be of advantage in
many ways. By immobilising the cells on porous carriers, cells are retained in the
system during continuous operation. This is of significance when cultivating cells,
which are adherent, have a high demand of substrate and slow growth-rates. This is also
accomplished by hollow-fibre modules or systems were cells are immobilised on or into
microcarriers, but these systems proved to be difficult in scale-up, a factor which is
especially important for the development of artificial organs, where a large amount of
cells is needed. A scale-up is easily achieved for the fixed-bed system by pumping the
medium radial through the fixed-bed instead of axial and therefore preventing the risk
of an oxygen limitation.

Due to the development of new immortalised hepatocyte lines that still show
considerable Cytochrome P450 activity, an approach to an artificial liver support
system might be accomplished with the aid of a bioreactor containing these cell lines
instead of primary hepatocytes which do not proliferate in such a system.

657
O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 657-659.

© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

658

3. Cell Line and Culture Conditions

The immortalised mouse hepatocyte line mHep-R1 was used for cultivation. It shows in
comparison to primary hepatocytes lower but still significant P450 activity of the
subenzymes CYP1A2/3A4 and CYP2C (Stange et al., 1995).

For growth low glucose Dulbecco’s MEM (Life Technologies, Germany) was used
as basal medium. This medium contains 25 mmol HEPES, 4 mmol L-glutamine
and 5.5 mmol glucose and was supplemented with 5% (v/w) foetal calf serum. The
glucose level was increased in the 40 ml fixed-bed system after the first week by
addition of glucose, because growth in preliminary experiments proved to be clearly
glucose limited.

For the batch experiments cultivation took place in 25 T-flasks. Growth curves
were obtained by analysing a doublet of flasks each day.

The 40 ml fixed-bed reactor, a modified reactor from meredos (Germany), contained
two DO sensors, automatic pH and temperature control (Pörtner et al., 1997). Cellulose
carriers (Cellsnow, Biomaterials, Japan) with a diameter of approx. 5 mm were used to
immobilise the cells.

4. Results

4.1. BATCH EXPERIMENTS

The cells grew to a maximum cell density
of cells after 72 hours with a
membrane intact index of approx. 95%.
The reason for cell death in the following
death phase was clearly due to glucose
limitation.

4.2. FIXED-BED REACTOR

Continuous operation of the fixed-bed
system was initiated 2 days after
inoculation. The dilution rate was kept
constant at 6.25 ml medium per ml fixed-
bed and day throughout the 75 days of
operation. Glucose was again the growth
limiting factor during the first few days

(Fig. 1). To improve growth conditions, the
glucose concentration in the feed was
raised to 16 mmol on the day,
resulting in increased glucose consumption
and lactate production rates
as well as a higher glucose concentration.
Several steady-states with different glucose

659

concentrations in the feed were obtained in the following 60 days. A limitation of

glucose can be excluded for feed concentrations over 16 mmol because no further

increase in the glucose consumption rate was observed. With a high
concentration of approx. 1.5 mmol , glutamine was also not limited (Fig. 2). We

assume that a substance in the serum is the cause for limited growth because the glucose
consumption and lactate production rates decreased rapidly when the serum level was
reduced to 4% in the feed on the day (Fig. 1) and started to recover on the day

after returning to the serum level of 5%. The average production rate was approx.
0.6 mmol for ammonia and 0.08 mmol for urea. 4.6 mmol ammonia
was added to the feed on the 30th day (Fig. 2) to estimate the stability of the production
rates at higher ammonia levels. The ammonia production rate decreased to
and the urea production rate to approx. 0.03 mmol

Several carriers were removed at the end of the experiment, and the total cell density
was determined with the crystal violet method to be cells per ml fixed-bed.
This corresponds to cells per ml carrier with an external porosity of 40%.

5. Scale-up considerations

For an adult human approx. 300 g of full
functional hepatocytes (or 20% of the
average liver mass) are able to maintain
sufficient function for detoxification. If
we assume that a permanent cell line still
shows 100% of the function of primary
hepatocytes, an extracorporeal liver
support system would need a fixed-bed
volume of approx. 6.6 litres. Large fixed-
bed reactors are operated by pumping the
medium in a radial way through the bed
rather than axially because of a possible
oxygen limitation in fixed-beds over 15
cm of height.

References

We thank Prof. Strauß of the Max Delbrück Center
for Molecular Medicine, Berlin, Germany, for the
donation of the mHep-R1 cell line and Biomaterials,
Japan for the supply of the Cellsnow carrier.

Acknowledgement

Pörtner, R., Rössing, S., Koop, M., Lüdemann, I. 1997.
Kinetic studies on hybridoma cells immobilised in fixed bed
reactors. Biotech. Bioeng. 55: 535-541
Stange, J., Milzner, S., Strauß, M., Fischer, U., Lindemann,
S., Peters, E., Holtz, M., Drewelow, B., Schmidt, R. 1995.
Primary or established liver cells for a hybrid liver?
Comparison of metabolic features. ASAIO J (United States)
43 (3):M10-M315

HIGH DENSITY PERFUSION CULTURE OF PRIMARY RAT
HEPATOCYTES FOR POTENTIAL USE AS A BIOARTIFICIAL LIVER
DEVICE

K. BRATCH, A.J. STRAIN & M. AL-RUBEAI
School of Chemical Engineering and Liver research laboratory,
University ofBirmingham, Edgbaston, Birmingham B15 2TT, U.K.

1. Abstract
Primary rat hepatocytes were cultured in a perfusion bioreactor system in an attempt to
develop a bioartificial liver support device. The bioreactor is based on immobilisation
of cells attached to microporous carriers packed into a stationary bed reactor. The
bioreactor is perfused with media and oxygenated independently thereby allowing
better homogeneity through optimum control of nutrient and oxygen transfer.
Hepatocytes which are essentially non-dividing cells when cultured in vitro, are
required to be maintained for prolonged periods with minimum loss of cell viability
and function. Attachment to microcarriers was optimised by investigating the influence
of various parameters including temperature, serum levels, pH and cell density.

Keywords: tissue engineering, hepatocytes, packed bed reactor, bioartificial liver.

2. Introduction
Fulminant hepatitis remains a highly lethal liver disease with the mortality rate
remaining very high in the absence of transplantation. Due to a shortage of donors
many of the patients will die while awaiting liver transplantation, therefore there is a
need to develop a bioartificial liver system to help keep these patients alive until either
an organ becomes available or the liver recovers from injury. The major problem in the
development of an artificial liver has been the complex biochemical nature of the
organ. An artificial support system should be capable of carrying out the complex
synthetic and metabolic functions of the liver as well as detoxification and excretion.

The most important functioning cells constituting the liver are hepatocytes; in the
case of bioartificial livers isolated hepatocytes are employed as the biological material.

We are currently using primary rat hepatocytes as the biological component for a
bioartificial device. Hepatocytes are anchorage dependant cells and the attachment of
hepatocytes to an extracellular matrix is essential for maintenance of function.
Hepatocytes are usually cultured as monolayers on collagen coated surfaces. In order to

661

O.-W. Merten et al. (eds.), New Developments and New Applications in Animal Cell Technology, 661-663.
© 1998 Kluwer Academic Publishers. Printed in the Netherlands.

662

obtain the high numbers of cells that are required to constitute a bioartificial liver it
would be advantageous to immobilise hepatocytes on microporous carriers. This would
provide a high surface to volume ratio thereby achieving a larger cell attachment area.
Unlike cell lines primary rat hepatocytes have a limited capacity to divide in vitro,
therefore large numbers of cells must initially attach to the microcarriers if they are to
be utilised in a liver assist device. As a result our work is aimed at investigating the
attachment of hepatocytes to microporous carriers (fibracell discs) and parameters
which may influence this are under investigation.

3. Materials and Methods
Hepatocytes are obtained from male Wistar rats by standard perfusion of the liver via
the hepatic portal vein with collagenase.The bioreactor consists of a glass cylinder

packed with microporous carriers (fibracell discs). Oxygen is supplied via a central
core of coiled silicone tubing held in position by a glass support. The fibracell discs
consist of a 50:50 mixture of polyester and polypropylene, with dimensions of 6mm in

diameter and 2mm in thickness.

Experiments to investigate various factors effecting the attachment of cells to fibracell
discs were carried out in duran bottles containing magnetic stirrers. The discs and cells
suspended in DMEM were stirred at 72rpm in a 37°C incubator.
MTT assays using MTT at a concentration of 5mg/ml were carried out on samples of
the discs and quantitative data was obtained by solubilising the formazan product with
0.04M HC1 in propan-2-ol and measuring absorbance at 570nm using a
spectrophotometer.

4. Results
The two plots in Figure 1 represent data from two separate experiments investigating
the effect of incubation time on cell attachment to fibracell discs. Greatest attachment
in both cases was observed after shorter incubation times.
When comparing attachment in the presence and absence of serum, the presence of
serum markedly increases cell attachment to the discs. However, it was apparent that
cell attachment was as efficient at low serum levels as higher serum levels (Figure 2).
The larger the ratio of cells to discs the more cell attachment observed (Figure 3), but
this did not appear to be linear, further experiments are being carried out to
investigate the saturation point of the discs.
The effect of pH on the attachment of cells to the microcarrier is shown in Figure 4.
The results indicate that greater attachment appears to be achieved at pH 7.3 with
decreasing attachment at higher pH.