Journal of Virological Methods, 8 (1984) 63-7 1 63

Elsevier

JVM 00288

A MICROCARRIER CELL CULTURE SYSTEM FOR LARGE SCALE

PRODUCTION OF HEPATITIS A VIRUS

ANDERS WIDELL, BENGT GdRAN HANSSON AND ERIK NORDENFELT

Section of Clinical Virology, Deparimeni of Medical Microbiology, University of Lund, MalmB General
Hospital, Maim& Sweden

(Accepted 25 October 1983)

Hepatitis A virus (HAV) was isolated from human faeces using a fetal rhesus monkey kidney cell line
(Frhk-4). Infectious medium from passage 12 was used to inoculate a large (5000 cm2) microcarrier cell
culture maintained in suspension. The microcarriers used were swollen, collagen-coated dextran beads on
which it was easy to propagate Frhk-4 cells. Intra- and extra-cellular virus levels were assayed and
compared with conventional cultures in 25 cm2 plastic flasks. The results show that virus production per cell
was similar in both systems. The number of cells per area unit in confluent cultures was initially lower in the
microcarrier culture but subsequently increased. Two to three weeks post inoculation the virus yield per
area unit in the microcarrier system was half of that of the conventional culture. The lower cell density per
area unit in the microcarrier system was compensated by the large growth area that could be maintained in a
single vessel and the total production of virus was substantial. Weekly harvests of medium with HAV
antigen titres around IO-* contained antigenic material sufficient for several thousands of anti-HAV IgM
tests. Propagation of HAV in microcarrier cell cultures thus seems a safe and simple way to produce large
amounts of HAV.

hepatitis A virus propagation microcarrier cell culture

INTRODUCTION

The first successful report of propagating hepatitis A virus (HAV) in vitro was
published in 1979 by Provost and Hilleman who used HAV, serially passaged in
marmosets, as inoculum. Soon afterwards the same group and others (Frosner et al.
1979; Flehmig, 1980) were able to grow HAV directly from faeces. Subsequently a
number of different cell lines have been used (Locarnini et al., 1981; Gauss-Miiller et
al., 1981). A typical feature in the propagation of HAV infected cell cultures has been
the absence of cytopathic effect. For this reason virus propagation must be monitored
by other means such as immunofluorescence or radioimmunoassay (RIA).
A new principle for expanding tissue cultures ofanchorage dependent cells was first
described by van Wezel in 1967 who used small spheres (microcarriers) kept in
suspension to increase the area available for cell growth. This technique has been
further developed and applied to different cell systems (Giard et al., 1977; Fohring et

Ol66-0934/84/$03.00 0 1984 Elsevier Science Publishers B.V.

64

al., 1980). This study described the propagation of hepatitis A virus in fetal rhesus
monkey kidney cells (Frhk-4) growing on cross linked dextran microcarriers and the
results are compared with conventional methods of propagation.

MATERIALS AND METHODS

A. Cell system and source of HAV

Culture vessels For conventional cell cultures, sterile, disposable 25 cm2 plasticflasks
were used (Nunclon Delta, Roskilde, Denmark).

Cells The fetal rhesus monkey kidney cell line (Frhk-4) was a kind gift from Dr.
Bertram Flehmig, Tubingen, F.R.G.. These ceils, at that time at the 90th passage level,
could be split 1:2 to 1:3 weekly. The cells were kept in closed vessels at 37°C during all
stages (including the period of virus propagation).

Medium Eagle’s Minimal Essential Medium with Hanks’ salts (Gibco, Paisley, U.K.)
was supplemented with 20 mM of Hepes (Gibco) and 3% of fetal calf serum (FCS)
(Flow Laboratories, Irvine, U.K.). During cell growth, FCS concentration was in-
creased to 10%. The medium also contained 250,000 IU of penicillin and 100 mg of
streptomycin/l.
Trypsin solution (0.25%) (Flow Laboratories, Irvine, U.K.) with 0.02% of EDTA
(Titriplex III, Merck, Darmstadt, F.R.G.) was prepared in Ca*’ and Mg*+ free phos-
phate buffered saline (PBS).

HAV inoculum HAV strain H 141 was isolated at our laboratory from the stools
collected from a patient with hepatitis A 2 days after the onset ofjaundice. The patient
had acquired the infection in India.
During the primary isolation cell bound virus was detected by RIA after 4wk. After 6
wk HAV was also detected in the medium. Positive cell homogenates were passed twice.
Subsequently, the medium was used for passage to new cultures once a week for
another nine times. This strain in passage 12 was used as the inoculum. One millilitre
of this material had an infectivity of lo6 TCID,, and an antigen titre by RIA of lo-t-“.

Detection of HAV A microtitre solid phase RIA as described elsewhere (Purcell et al.,
1976) was used. Briefly, test samples were incubated in the wells of a microtitre plate
previously coated with human anti-HAV. After washing, ‘2SI-labelled anti-HAV-IgG
was added and the bound radioactivity measured. The mean of 5 negative controls(N)
was calculated and HAV reactivity expressed as the ratio between the individual
sample (S) and N. S/N ratios above 2.1 were regarded as positive. In each test run, one
HAV positive sample of faeces was included as a positive control. Test specificity has
been confirmed by blocking experiments using pre-immunization and convalescent
 65

sera from an experimentally infected chimpanzee (kindly provided by Dr. R. Purcell).

Cell bound HA V After removal of medium, a standard volume of PBS solution was
added and the material was then freeze-thawed three times to -20°C before testing.

Supernatant HA V The medium was centrifuged at 2000 X g for 10 min and the clear
supernatant used for testing.
The titration curve for cell bound virus closely resembled that of HAV in faeces
serially diluted in PBS. However, when serial dilutions of the medium were made in
PBS and tested by HAV RIA, a prozone effect occurred (Fig. 1). This finding made
end point titrations necessary in all cases where quantitative determinations were
desired.

B. Establishment of a microcarrier cell culture

Culture vessel For the suspension cultures a specially designed siliconized culture
vessel (F 7607, Techne, Cambridge, U.K.) with a magnetic driving unit (MC!+104,
Techne, Cambridge, U.K.) was used. Four l-l vessels, each with a working volume of
500 ml, could be stirred simultaneously.

,
la+ io-"5 10-l 10-'~5 l(r' ~O%ILUTION

Fig. 1. Titrations of tissue culture HAV in PBS determined by RIA. Cell-bound virus from a con-
ventional (0 -0) and a microcarrier culture (0 -0) compared with supernatant virus from aconven-
tional (o---o) and a microcarrier culture (O---O). S/N ratios Z 2.1 were regarded as positive.
66

Microcarriers Cytodex 3 (Pharmacia Fine Chemicals, Uppsala, Sweden) a recently

developed type of collagen-coated dextran microcarrier was used. When added to
medium, the dextran beads swell to an average diameter of 175 urn. The volume of the
beads increase about 15 fold and the surface area from 1 g of dry Cytodex 3 becomes
4500 cm2 (Pharmacia Fine Chemicals, 1981). Swelling and preparation of microcar-
riers and cell attachment were performed according to the recommendations by the
manufacturer with some modifications.

Preparation of microcarriers 0.3 g of dry Cytodex 3 was allowed to swell overnight at

20°C in 100 ml of PBS using a siliconized vessel. The buffer was then changed once to
fresh PBS and the microcarriers sterilized at 120°C for 15 min. After removal of the
buffer, the beads were washed once with growth medium containing 10% FCS and
resuspended in 60 ml of growth medium at 37°C for immediate use.

Attachmentofcelfs Largenumbersofcells(30-40millions)wereobtainedaftertrypsin
treatment of conventional monolayer cultures in flasks. The cells were suspended in 30
ml of trypsin solution and then immediately transferred to the microcarriers at 37”C,
stirred (25 RPM) for 1 min and then allowed to settle. The mixture was stirred for 1
min each at the 30th min during the first 3 h, then by automatic control for 30 s each
5th min for 3 h, and thereafter continuous stirring started. After 24 h the volume of
medium was increased to about 250 ml. Growth medium was changed twice weekly
until full confluence of cells on the microcarriers appeared. Subsequently, mainte-
nance medium with 3% FCS was used and changed once weekly. Before changing the
medium, the stirring was interrupted, whereupon microcarriers sedimented fully
within l-2 minutes. Only half of the medium volume was changed each time.

Propagation of cells on microcarriers The 250 ml culture could be further expanded to a

500 ml culture by using the cells on the microcarriers. All medium was then carefully
removed and the culture rinsed twice with sterile PBS. Fifty ml of trypsin solution was
added and after cell detachment the cells were easily separated from the microcarrier
beads by filtering through a 0.1 mm nylon grid. The cells were then added to 2-3 times
larger quantity of prepared microcarriers than the original one, and the procedure
repeated as described above.

Design of study Propagation of HAV in a microcarrier cell culture system wascarried

out in parallel with propagation in conventional flask cultures. In order to compare
the two systems, the growth area, i.e. the area occupied by confluent ceils, was chosen
as the standard to which virus inoculation and harvests were related. In a confluent
culture based on 1.1 g of Cytodex 3 in 500 ml of medium, the growth area is 5000:500 =
10 cm2 per millilitre of medium. Two such cultures and seven 25 cm* flask cultures
with 5 ml of medium each were prepared. Samples of 2.5 ml, drawn from well
suspended microcarrier cell cultures, thus had the same growth area as the 25 cm2
 67

flasks. In both systems, the age of the cells since last passage was 2 wk at the time of
virus inoculation,

Cell counting The cells of two 2.5 ml portions of each suspension culture and one 25
cm2 flask were treated with trypsin until all cells had detached and no aggregates were
seen. Detached cells were stained with trypan blue and counted in a haemocytometer.
Cell counting could only be performed at the beginning of each experiment since older
cells aggregated after trypsin treatment.

Virus inoculation of cell cultures Before inoculation, the medium volumes in both the
suspension cultures and the flasks were reduced to one-fifth. Five millilitres of the
virus inoculum were added to one 5000 cm* culture and 25 ul to each of six 25 cm2
flasks, i.e. the same amount of virus per growth area unit was added to each culture.
No inoculum was added to the remaining 5000 cm2 culture which served as a negative
control. After 4 h the volumes of medium were adjusted to normal (500 ml and 5 ml,
respectively). The suspension cultures were stirred at 25 rpm both during and after
inoculation whereas the 25 cm2 flasks were kept static.

Harvesting Two 2.5 ml ~01s. of the infected and one 2.5 ml portion of the uninfected
suspension culture and all medium and cells from one flask were harvested twice
weekly for 3 wk. Medium was removed from both suspension culture samples after
sedimentation and from the flask, centrifuged at 2000 X g for 10 min and stored at
-20°C until tested for virus in the supernatant. 2.5 ml of PBS were added to the cells of
both the microcarrier samples and the flasks, freeze-thawed three times and stored at
-20°C until tested for cell-bound virus.

Examination of suspension cultures Infected and uninfected suspension cultures were

examined twice a week for signs of cytopathic effect using a conventional light
microscope. The pH of the cultures was determined either by observing the colour
indicator of the medium or by using a pH-meter.

RESULTS

General observations at the establishment of Frhk-4 cell culture on Cytodex microcar-

riers
In all trypsin treatment steps it was essential to use cultures at the end of exponential
growth. Cultures older than 2 wk were unsuitable for cell passages since such cells
were difficult to detach and they rapidly formed large aggregates with subsequent cell
loss.
The majority of the cells attached to the microcarriers within 6-8 h. At this stage, it
was easy to count the number of cells on the microcarrier beads (Fig. 2a). Attachment
of lo-15 cells per bead usually led to confluent cell growth within a few days, whereas
Fig. 2. Appearance of Frhk-4x11s on Cytodex 3 microcarners by light mnoscopy. (2a) Typical appearance
6 h after cells have been added to microcarriers: (2b) the confluent culture at virus inoculation; (2~) the same
culture 21 days later.

approximatively 5 cells per bead in general led to confluent growth on most microcar-
riers but not until 2-3 wk later. Microcarrier beads with unattached or with only single
cells remained empty.

Cell numbers
Although the cells were fully confluent on 95% of the microcarriers at the time of
virus inoculation a difference in cell density between conventional flasks and micro-
carrier samples was observed. While one 25 cm* flask contained approximately lo6
cells (40,000 cells/cm*), the corresponding 2.5 ml of suspension cultures contained
only 1.35 . lo5 cells (5400 cells/cm2) and 1.20. lo5 (4800 cells/cm*) at that time. There
was a variation in the number of cells on different microcarrier beads. On more than
half the beads the cells were large and rather thin (Fig. 2b). During the weeks following
the virus inoculation this difference disappeared when the cell numbers increased as
observed by light microscopy (Fig. 2~).

Propagation of HAV in conventional and microcarrier cell cultures

During the 3 wk of observation no cytopathic effects or pH changes were seen in the
infected cultures compared to the uninfected control culture. At the time of the first
harvest (4 days post inoculation) HAV was detected by RIA in both cell homogenate
 69

and medium from the microcarrier culture as well as from the flask culture. Quantita-
tion by end point titrations was carried out on all samples harvested at day 7, 14 and 2 1
using serial loo.5 dilutions. The end point titre chosen was the point where the titration
curve intercepted the 2.1 cut off level (Fig. 1).
The level of cell bound virus in the conventional flask was maximal at day 7 post
inoculation (Fig. 3). At that time cell bound virus from an identical growth area of the
microcarrier culture homogenized in 2.5 ml of PBS was about 8 times (10°.9) lower.
This lower cell bound virus harvest was found in a system that at the time of virus
inoculation contained about eight times fewer cells per area unit.
The eight-fold difference of cell bound virus per area unit between the systems at
day 7, subsequently diminished to about two-fold ( 10°.2-100.3) in parallel with increa-
sing cell density on the microcarrier beads observed by light microscopy.
The increase of virus in the medium parallelled the increase of cell bound virus and
in both systems the end point titres reached the 10e2 level. In the microcarrier system
the medium supply was one-fourth of that of the conventional system (0.1 ml/cm2
versus 0.2 ml/cm2 and only half the volume of medium exchanged weekly).

DISCUSSION

In order to obtain large amounts of HAV grown in tissue culture, it is necessary to

have a suitable cell type and a suitable virus inoculum. HAV propagates in cell culture
without cytopathic effect, often very slowly initially and in most systems without virus
release into the medium (Frosner et al., 1979; Gauss-Miiller et al., 1981; Locarnini et
al., 1981). On the other hand, Flehmig (1980) found virus release into the medium
using Frhk-4 cells and human fibroblasts (Flehmig et al., 1981). By performing
multiple passages (approximately 50 passages) using media as inocula and decreasing

HAV TITRE

10"

lO'.O

Fig. 3. HAV end point titres determined by RIA in material harvested 7, 14 and 21 days post inoculation.
Cell-bound (a- o) and supernatant (o---o) virus from conventional flasks compared with cell-bound
P- q) and supernatant (O---O) virus from the microcarrier culture.
70

the passage time to 1 wk it has been possible to obtain fast growing variants of HAV
with both high antigen and infectivity levels in the medium (B. Flehmig, pers. comm.)
The same principle was therefore applied to our H 141 strain resulting at passage 12in
highly infectious medium (TCID,, around 10e6) and HAV antigen level of approa-
ching lo-’ 1 wk post inoculation as measured by RIA.
Recently it has been found that the Frhk-4 and Frhk-6 cell lines have been infected
with a polyoma virus, probably of bovine origin, at lower passages (Parry et al., 1983).
It is not yet clear if this coinfection has any influence on HAV replication and
excretion. However, the specificity of the HAV test is not affected since all uninfected
controls were negative by the HAV RIA test.
Various types of microcarriers have been used to obtain large numbers of cells for
different purposes. Microcarrier cell cultures have been used successfully for produc-
tion of viruses like poliomyelitis (van Wezel et al., 1978; Mered et al., 1980), rabies
(van Wezel et al., 1978) and foot-and-mouth disease virus (Meignier et al., 1980). In
these studies, the suitability of microcarrier cell cultures for vaccine production has
been demonstrated, since large cultures can easily be maintained in fermentors.
For HAV production in a microcarrier system it was necessary to establish suitable
cells on microcarriers. This could be accomplished readily with Frhk-4 cells on
dextran beads. However, at the time of confluent growth, the number of cells per area
unit was lower in the microcarrier culture than in the conventional flasks. Lower cell
density in microcarrier systems has also been described by Mered et al. (1980) who
found 2- to 3-fold lower cell density in a microcarrier system. This is probably due to
the cell adherence on the different surfaces and materials (round versus flat, dextran
versus plastic). Control of pH which is considered to be a major factor for successful
microcarrier systems (Fiihring et al., 1980) was satisfactorily accomplished when
Hepes (20 mM/l) was included in the medium.
The cell bound virus yield per area unit 7 days post inoculation was 8 times lower in
the microcarrier system than in the plastic flask. This corresponds to the differences in
cell density per area unit (up to 8 times initially). The virus production per cell was thus
similar in the two systems. The subsequent increase of cell bound virus in the
microcarrier system can partly be explained by increasing cell density. Antigen titres
in the media parallelled the cell bound virus levels and eventually in both systems
reached just above lo-*.
The production of virus in this microcarrier system is substantial. The weekly
harvest of medium (250 ml) from week 2 onwards, contained antigenic material
sufficient for approximately 15,000 anti-HAV IgM tests by RIA (Hansson et al., 1981)
of the kind which is used in our laboratory. Repeated harvests of media can be
performed for long periods, months, perhaps even years. One tube of cells from the
primary isolation of the H 141 strain has been maintained for 23 mth in our laboratory
continuously producing HAV (data not shown).
One definite advantage of the microcarrier cell system is the simple handling when
media are changed. Contaminated glassware is kept at a minimum thus decreasing the
 71

risk for laboratory personnel. Handling of one single vessel also minimizes the risk of
contamination from the outside. Another advantage is the possibility of frequent
sampling of aliquots for monitoring of cell bound virus levels. Thus, large numbers of
vessels inoculated simultaneously, but designed to be harvested at different times, are
avoided.
Propagation of HAV in microcarrier cell system seems a safe, cheap and practical
way to obtain virus for diagnostic and research purposes. It is likely that such cultures
could be used for production of future HAV vaccines.

ACKNOWLEDGEMENTS

This work was supported in part by a grant from Alfred osterlunds stiftelse. The
author is grateful to Mrs. Agneta Nilsson, Miss Carina Sjoberg and Miss Ann Sofi
Persson for skilful technical assistance.

REFERENCES

Flehmig, B., 1980, Med. Microbial. Immunol. 168, 239.

Flehmig, B., A. Vallbracht and G. Wurster, 1981, Med. Microbial. Immunol. 170, 83.
Fdhring, B., S.T. Tjia, W.M. Zenke, G. Sauer and W. Doerfler, 1980, Proc. Sot. Exp. Biol. Med. 164. 222.
Frosner, G.G., F. Demhardt, R. Scheid, V. Gauss-Miiller, N. Holmes, V. Messelberger, G. Siegl and J.J.
Alexander, 1979, Infection 7, 303.
Gauss-Miiller, V., G.G. Frosner and F. Deinhardt, 1981, J. Med. Viral. 7, 233.
Giard, D.J., W.G. Thilly, D.I.C. Wang and D.W. Levine, 1977, Appl. Environm. Microbial. 34, 668.
Hansson, B.G., J.K. Calhoun, D.C. Wang, S.M. Feinstone, R.H. Purcel1,C.S. Pannuti, J.L. Pereira, R.S.
Koff, J.L. Dienstag and S. Iwarson, 1981, Stand. J. Infect. Dis. 13, 5.
Locarnini, S.A., A.G. Coulepsis, E.G. Westaway and I.D. Gust, 1961, J. Viral. 37, 216.
Meignier, B., H. Mougeot and H. Favre, 1980, Dev. Biol. Stand. 46, 249.
Mered, B., P. Albrecht and H.E. Hopps, 1980, In Vitro 16, 859.
Parry, J.V., J.E. Richmond and S.D. Gardner, 1963, Lancet I, 994.
Pharmacia Fine Chemicals, 1981. In: Microcarrier Cell Culture, Principles and Methods. (Almqvist and
Wiksell Tryckeri AB, Uppsala) p, 125.
Provost, P.J. and M.R. Hilleman, 1979, Proc. Sot. Exp. Biol. Med. 160, 213.
Purcell, R.H., D.C. Wang, Y. Moritsugu, J.L. Dienstag, J.A. Routenberg and J.D. Boggs, 1976, J. Immu-
nol. 116, 349.
Van Wezel. A.L.. 1967, Nature (London) 214, 64.
Van Wezel. A.L.. G. van Steenis, Ch.A. Hannik and H. Cohen, 1978, Dev. Biol. Stand, 41. 159.