HYBRIDOMA Volume 9, Number 2, 1990 Mary Ann Liebert, Inc., Publishers
Loss of Antibody Productivity During Long-Term
Cultivation of a Hybridoma Cell Line in Low Serum and Serum-Free Media SADETTIN S. OZTURK and BERNHARD
0.
Department of Chemical Engineering, University of Michigan,
PALSSON
Ann Arbor, MI
ABSTRACT A murine hybridoma cell line (167.4G5.3) was adapted to grow in different serum concentrations over a six month period of time. Adaptation to low serum and to serum-free media improved growth rates, but at'low serum (1.25%) the antibody productivity was diminished. Flow cytometric analysis showed the presence of two distinct cell populations with respect to intracellular and surface antibody concentrations. The loss in antibody productivity during adaptation could be attributed to the appearance of a low antibody-containing cell population. Cultures maintained at high serum concentrations did not loose the original high antibody productivity. In a separate experiment the kinetics of growth improvement and loss of antibody production were studied for adaptation from 5% to 1.25% serum-containing media. Over a time period of about four months, the population shifted completely from high-producing cells to low-producing ones in response to the 1.25% environment. A shift-up from 1.25% to 20% serum resulted in the elimination of the low producing population. These results suggest that, for the cell line used, serum-containing factors prevent the loss of antibody productivity.
INTRODUCTION
Hybridoma cell technology is widely used for the preparation of monoclonal antibodies (MAbs) wide spectrum of antigens. The method of establishing permanent cell lines capable of producing antibodies directed to predefined immunogen is based on the fusion of immune lymphocytes with myeloma cells adapted for growth in tissue culture conditions (Esshar, 1985). Several genetic processes occur during and after cell fusion. During cell fusion, the genes from lymphocyte and myeloma fusion partners are rearranged to generate the set of chromosomes that are retained in the resulting hybridoma cell. The genes may continue to be rearranged after the fusion, and chromosomal losses have been reported over longer periods of time (Hunkeler ei al, 1989). The appearance of a low producing hybridoma population in long-term cultures has been observed experimentally (Frame ei al, 1989; Heath et al, 1989). This genetic instability is very important to the economics of long-term hybridoma cell culture (Dean, 1989). Frame et al (1989) showed a decrease with time in the overall antibody production rate that was calculated from the to
a
antibody concentrations and the total viable cell counts in a chemostat. Their flow cytometric analysis revealed heterogeneity of the cell population: a fraction of the cells were lacking the antibody (non-producing cells), while the rest of the cells were producing at a constant specific (per cell) production rate (producing cells). The loss in the observed antibody production rate 167
attributed to the increase in the size of non-producing cell population. A similar observation reported by Heath et al (1989). The non-producing cell population size increased in time and the specific antibody production rate decreased correspondingly. Cell culture media typically contain serum to support cell growth. Although growth is greatly enhanced in the presence of serum, economical and processing considerations favor the use of low serum, or serum-free media for large-scale cultivation. The cost of serum, lot to lot variations in serum quality, undefined composition, and the complications in antibody purification that results from the presence of serum protein have stimulated research on the development of serum-free media. Several serum-free formulations have been developed and are used successfully for a variety of mammalian cells. Antibody productivities in serum-free media were reported to be comparable to those obtained in serum-containing media (Murakami, 1989). Adaptation is needed to obtain improved cell growth in low serum or serum-free media (Wolpe, 1984). Cells have to be passed for several generations in the low, or zero, serum-containing media to achieve good cell viability. The number of generations needed for adaption depends on the cell type and the media used; some cells adapt after a few generations, some cells never adapt (Kovar, 1989). Success of the adaptation of cells to lower serum-containing media can be improved in a 'weaning' procedure in which the serum concentration in the passing media is decreased gradually was was
(Wolpe, 1984).
Several processes may take place during adaptation. Improvement in the growth rate is one of the changes that occur with adaptation to low serum-containing media. The improvement in growth rate suggests changes in the cells' response to any growth-promoting factors provided by serum. For instance, it has been shown that for human breast cancer cell line MCF-7 the number of estrogen receptors vary by the adaptation process (Briand and Lykkesfeldt, 1984). Elsewhere, we have reported that adaptation of hybridoma cells to low serum did not alter key metabolic rates and yield parameters (Ozturk and Palsson, 1989a). Although hybridoma metabolism is insensitive to adaptation to low serum levels, adaptation may alter the specific antibody productivity of hybridoma cells. There is very little information in the literature in this respect. In the present paper, we focus on changes in the antibody production rate that occur during adaptation and present a study of the kinetics of the loss of specific antibody productivity.
MATERIALS and METHODS
line, medium and culture maintenance The murine hybridoma cell line, 167.4G5.3, was kindly provided by Dr. J. Latham Claflin from the University of Michigan Medical Center. The antibody produced by this cell line is an IgGi, directed against the phosphorylcholine (Briles ei al, 1984). Hybridoma cells were made by fusion of BALB/c spleen cells with the nonsecreting plasmacytoma fusion line P3X63-Ag8.653. Antibody was generated from mice immunized with PC-keyhole limpet henemocyanin (KLH). The cells were maintained in 75 cm2 plastic T-flasks (Bélico Glass, Inc., Vineland, NJ) in Iscove's Modified Dulbecco's Medium (IMDM, Gibco Laboratories, Grand Island, NY) containing Fetal Bovine Serum (FBS, Gibco) supplemented with 100 Units/ml potassium penicillin G, and 100 /Jg/ml streptomycin sulphate (Sigma Chemical, St. Louis, MO). The cells were kept at 37° C and 5% COi atmosphere in humidified incubators (VWR Scientific, San Fransisco, CA). Cell
\
Adaptation of cells to low serum levels Cells were adapted to grow in different serum
concentrations by a stepwise procedure that lasted for six months. The cells were initially growing in IMDM with 20% serum (which we designate as 20%-cells). A sample of these cells were frozen and the rest were adapted in 10% serum by passing the cells in IMDM containing 10% serum (10%-cells). After two months of passage, the cells were passed in 5% serum (designated as 5%-cells). Similarly cells were adapated in 2.5% (designated as 2.5%-cells), in 1.25% serum (designated as 1.25%-cells), and in serum-free media (OPTI-MEM, Gibco). Cells were also adapted in one step to grow in 5% and 1.25% serum: exponentially growing 20%-cells were inoculated directly to 5% and 1.25% serum-containing IMDM. The success of this direct adaptation was dependent on the viability of the 20%-cells. Also, the removal of spent media increased the success of the adaptation. Similarly, cells were adapted to 1.25% serum from 5%. In this case, the course of adaptation was followed very closely. Cells were passed every two days and counted using a hemacytometer. The cells were then passed by a dilution ratio determined from the viable cell count. In this procedure, the cell concentration after each passage was kept at 4 104 viable cells/ml. The cell supernatants were saved for the analysis of metabolite and antibody concentrations. -
168
Effect of adaptation
on
cell
physiology
The frozen adapted cells were thawed and propagated in the incubator using T-flasks for batch kinetic study of growth, metabolism, and antibody production rate. Variable passing ratios were applied in order to get all the cells to high concentrations and into exponential phase of growth. The changes in cell behavior were examined in different serum concentrations. Each adapted %-cell was inoculated into IMDM containing different serum concentrations. Four different serum levels were used for this purpose: 1.25, 2.5, 5, and 10%. Each adapted %-cell was centrifuged at 1000 rpm (200 g) for 10 min to remove the spent medium. The cells were washed with fresh media with IMDM containing 1.25% serum and finally, they were distributed into spinner flasks containing different serum concentrations. Experiments were run in duplicates in 100 cm3 spinner flasks (Bélico Glass, Inc., Vineland, NJ) with a 50 cm3 working volume. The flasks were placed on magnetic stirrers (Bélico) in a humidified CO* incubator at 37°C with 5% COi. The agitation rate for the spinner flaks was at 100 rpm. A one milliliter sample was taken twice daily during exponential growth, but once per day during the death phase. After performing cell counts, the samples were centrifuged and the supernatants were stored at -80"C for subsequent determination of metabolite and monoclonal antibody concentrations. Antibody, IgGi, was quantified using an enzyme linked immunosorbent assay (ELISA) as described by Ozturk and Palsson (1989b). a
Flow
Cytometry
A Coulter EPICS 751 flow cytometer was used to analyze the antibody content of the cells. Forward angle light scatter (FALS) and 90" light scatter (90LS) data was used for the estimation of cell size and cell complexity. Viable cells could be differentiated from the dead cells by the differences in their FALS and 90LS characteristics. Non-viable cells were eliminated by gating. Cellular composition of the cells was determined by using fluorescent dyes. Propidium Iodide (PI) was used for the measurement of DNA and RNA content of the cells. Stock solutions of propidium iodide were prepared from powder (Sigma) as 1 mg/ml level using phosphate buffer saline (PBS) and stored in 1 ml aliquots at -80°C Cells were treated with RNase (Sigma) to cleave the RNA for the measurement of DNA content and DNA distribution of the cells. Likewise, DNase (Sigma) was used for the RNA analysis. Stock solutions of DNase or RNase were prepared in PBS as 1 mg/ml and stored in 1 ml aliquots at -80°C. Intracellular protein was stained with Fluorescein Isothiocynate (FITC, Sigma). The stock solutions of FITC were prepared as 1 mg/ml and stored in 1 ml aliquots at -80°C. Cells were stained for the surface and total intracellular IgGi by Fluorescein Isothiocynate conjugated goat anti-mouse IgGi (GAM-FITC, Southern Biotech). The stock concentration of GAM-FITC was 0.02 mg/ml. Cells were fixed with ethanol for analysis of DNA, RNA, intracellular IgGi content. Only, surface IgGi determination was carried out with intact cells. The cells were fixated as follows: cell samples were spun at 1000 rpm for 5 min and the cell culture media was removed. Ceils were then washed with PBS at pH 7.4 and spun down again for 5 min. The cell pellet was resuspended in 70% ethanol and the suspension was kept at A°C overnight. For staining, these fixed cells were spun down at 1000 rpm for 5 min and washed with PBS. The cells were then centrifuged for 5 min and the pellet was suspended in 200 fil GAM-FITC solution. The cells were incubated for 45 min at 4°C. Then the cells were washed with PBS and stained with a mixture of PI and RNase (for DNA determination) or DNase (for RNA determination). The stock solutions of PI and RNase (or DNase) were diluted in PBS to give a PI concentration of 20 fig /ml and a RNase (or DNase) concentration of 40 pg/ml. This double staining was carried out at 37"C for 30 min. The cells were finally washed with PBS and run in the flow cytometer. Protein content was measured separately with fixed cells. The protocol was the same for the measurement of intracellular IgG decribed above, with the exception that FITC was used instead of GAM-FITC. The surface antibody was measured with intact cells (no fixation), as mentioned above. The cell samples were washed with the PBS and stained directly with GAM-FITC for 45 min at i°C. The cells were then washed with PBS and run in the flow cytometer. RESULTS
Effect of adaptation
on
the cell
growth
and monoclonal antibody
production
The short-term response of adapted cells was assessed by inoculating them into batches containing different serum concentration and they were grown in batch culture for a period of two weeks. The growth rates were evaluated in the exponential growth phase and are presented in Figure 1A as a function of serum concentration (short-term) used in these batches. For all
169
the
rate increased by serum concentration. Qualitatively, the concentration was the same for all the adapted cells and was in agreement with the data of Dalili and Ollis (1989) and of Ozturk and Palsson (1989b): the effect of serum on the growth rate was pronounced at low concentration but a constant growth rate was reached at high serum levels. However, quantitatively there was a difference in the responses of adapted cells. Although all the cells originated from the same lot (20%-cells), after adaptation, they resulted in different growth rates for a given short-term serum concentration. All the cells grew slowly at low serum concentrations, i.e. 1.25% FBS. At this serum concentration, the 1.25%-cells showed the highest growth rate. The data shows that the cells acquire growth advantage by adaptation to lower serum concentrations. At any serum concentration, the 1.25%-cells and 2.5% cells grew most rapidly, followed by 5%-cells, and so forth. The growth rate of 1.25%-cells was almost twice that of the original 20%-cells in 1.25% serum. We have seen similar growth rate improvements in another hybridoma cell line by adaptation (Ozturk and Palsson, 1989a). Similar studies have been reported for human breast cancer cell line MCF-7 (Briand and Lykkesfeldt, 1984), and for various hybridoma cells (Kovar, 1989).
adapted cells, the specific growth
response of cells to
new serum
(B)
u II S.Ï i
¡3 Adaptation
Écrum,
%
S
1*
Adaptation
IS
2»
25
«mm, *t/t
Figure 1. Response of adapted cdla to différent aerum concentration» after adaptation. (A) apecific growth rates as a function of aerum concentration uaed in batch growth for different adaptations, (B) apecific antibody production ratea as a function of adaptation aerum levels in different aerum concentrations uaed in batch growth. The antibody production rates for the adapted cells are presented in Figure IB as a function of concentration used for adaptation. The specific antibody production rates were evaluated at different serum concentration for the adapted cells in the batch growth experiment. This plot summarizes both long-term (adaptation) and short-term effect of serum on the antibody productivity of the cells. The specific production rate was not influenced by the short-term exposure to serum levels (response to different serum levels indicated by different symbols in Figure IB). However, there is clearly an influence of adaptation on antibody productivity. The 5%-, 10%-, and 20%-cells showed essentially the same specific antibody productivity, whereas significant losses in productivity were observed for cells adapted to 1.25% and 2.5% serum. These cells did not grow in 1MDM without serum. However, they did grow in OPTI-MEM serum-free media. Although OPTI-MEM is not a serum-free reference point for serum-containing IMDM media, it is instructive to see that the trends observed with reducing serum are continued to serum-free OPTI-MEM. The same loss in antibody productivity was observed when the cells were adapted to OPTI-MEM serum-free media. The gross macromolecular composition of the cells was measured. Figure 2 shows the data on cell size (FALS), DNA, RNA, gross macromolecular protein and intracelluler IgGi content. The data is presented as histograms or distribution functions in which the channel number corresponds to the intensity of the property considered and frequency corresponds to the cell number. The cells were in the exponential growth phase at the time of analysis. The cells were almost identical in terms of size, DNA, RNA, and protein content (Figure 2). There was, however, a significant difference in the antibody content. The 5%-, 10%- and 20%-cells have essentially identical intracellular IgGi levels. The 1.25%- and OPTI-MEM-cells, on the other hand, had a second lower peak along with the same peak exhibited by the cells adapted to higher serum concentrations.
serum
170
Size
DNA
RNA
protein
log(IgG)
OPTI-MEM CmMmei namber
Figure 2. Flow cytometric data on tbe cellular composition of the adapted ceDa. Cell aise, DNA, RNA, and protein content of the cells were almost identical. However, the intracelhilar IgGi content was different in the celia adapted in 1.25%- and in OPTI-MEM-cella.
The appearance of two peaks is further demonstrated in Figure 3. In this two-parameter diagram, the cells are compared with respect to the antibody content and DNA structure. The original cells shown in panel A exhibit a single population. The cells have antibody in all stages of the cell cycle. The adapted cells (1.25%-cells) have the same population and a similar, but lower, antibody-containing population. This second population of cells was very similar in terms of IgGi distribution during the cell cycle and in their cell cycle behavior. contour
i
i
y
12 1Ü
ADAPTED (L23ft FBS)
ORIGINAL (20% FBS)
e-
:i
as
S
! 3d
l
4 3
t >•
is
....Q^£6seSfei >r
...o...%
.< >.
o-
o
..s*.
IT
a....«.....«...*..
»s.
ras-1
1.25*
i
1»
-
/.!• •«
i:iy»~ 1.«
4*
St
6»
!••
121
I
2«
A. Cell
growth during adaptation
rate can be
evaluated from:
where
is the
to
1.25% ~
serum, and B.
Figure
7. The intracellular
in cell number observed
antibody
content
corresponding antibody lévela.
The
growth
ln(i/To) 2
_
x/x0
»•
Time, br
11
expansion
«t
4*
Time, days
•
• •
1J
Figure 6.
single
^g^^^ ^ .
3.« 2J
a
aemm.
days
over
measured
the
two-day period between aubculturing the cells.
during adaptation
173
to
1.25%
aerum
(Figure 6).
that the initial cell concentration after each passage was constant at 4 104 cells/ml. The cells in 5% serum were also passed under the same protocol as a control. Figure 6 shows the increase in cell count during each two-day passing cycle. The growth rate is related to this ratio by fi ln(i/i0)/2 days. After being introduced to 1.25% serum, the cells doubled every 2 days for about a 20-day period. Then, the growth rate increased, and over a 3-month period, the growth rate reached the growth rate observed in 5% serum. The antibody production rate and the intracellular antibody content was followed using ELISA (Figure 6B) and flow cytometry (Figure 7). The cells in 1.25% serum began losing their antibody level after one month, and the fraction of low producers began to increase. •
—
i
—— day,
Figure 8. (A) The fraction of high producers and (B) to
4*
the overall
antibody production rates during the adaptation
aerum.
DISCUSSION The present study has expanded our knowledge of the kinetics of the loss of antibody productivity in long term hybridoma cell cultures and how it is influenced by serum. When 167.4G5.3 hybridoma cells are adapted over a long period of time to low serum-containing media or serumfree media, the cells lose their antibody productivity. The loss of antibody productivity may be directly linked to the appearance of low-producing cells. For a mixed population of high- and low-producers, the apparent antibody production rate will be a cell-density weighed-average of the specific production rates of the high and low producers. The apparent antibody production rate would be equal to the specific antibody production rate of 5%-, 10%- and 20%-cells only when the fraction of low producers is zero. The production rate of the high producer is expected to be qn = 0.2 pg/cell/hr as it is in the 5%- and other high serum-adapted cells. The production rate of low producers is qr, and if the fraction of the low producers is / then the overall antibody production rate is: q
=
?/f(l-/) + /?£
Then, if the fraction of low producers is high, the observed antibody production will be low. Antibody productivities were calculated using the cell counts and the antibody concentrations given in Figure 6 for cells in 5% and 1.25% serum and are presented in Figure 8B. This overall antibody production rate was constant for the cells kept in 5% serum (control). However, the overall productivity decreased over time in 1.25% serum culture. This decrease was in parallel to the decrease in the fraction of high producers in 1.25% serum determined from Figure 7 and plotted in Figure 8A. Hence as the above equation predicts, we see a decrease in the overall antibody production rates that corresponds to the increasing number of low antibody containing cells in the culture. The loss in antibody productivity is thus directly related to the appearance
of the cells with a low antibody content. There are three possible hypotheses to explain our observations. Firstly, we could be observing the appearance of a second clone in the culture, created by genetic drift, that has lost or has reduced ability to produce antibody. The observations we report thus are, simply, the dynamics of a two-population system where the cell composition of the culture shifts with culture conditions. If this hypothesis is true, our results indicate that higher levels of serum support the growth of a high producing clone, whereas long-term adaptation to lower serum-containing media favors the growth of a low- or non-producing clone. A second hypothesis would relate to regulation of 174
transcription or translation rate. In this case, one would hypothesize that the cells gradually change their surface receptor portfolio with time, and this slow change would then, in turn, alter the stability of mRNA or the expression rate of the antibody-coding genes. A third possibility is that the cells might be producing an antibody molecule that is not IgGi and is thus not detected by our assays. Our present data cannot conclusively determine which mechanism might be at the
work. ACKNOWLEDGEMENTS This work was supported by National Science Foundation Grant No. EET-8712756. The authors thank Dr. J. Latham Claflin for providing the hybridoma cell line used in this study and Mark Cameron for his assistance with flow cytometry.
REFERENCES 1.
2. 3. 4.
5. 6.
7.
Esshar, Z., (1985) "Monoclonal Antibody Strategy and Technique," In: Hybridoma Technology in the Biosciences and Medicine, T. A. Springer, Ed. Plenum Press, New York,
3-41, 1985. Hunkeler, N., Hu, VV-S., and Srienc, F., (1989) "Hybridoma Subpoulations Affecting Culture Productivity," In Proceedings of the Annual American Chemical Society Meeting. Frame, K, Sen, S., and Hu, W-S., (1989), "The loss of Antibody Productivity in Continuous Culture of Hybridoma Cells," Biotech. Bioeng., accepted. Heath, C. A., Dilwith, R., and Beifort, G. (1989) "Methods for Increasing Antibody Production in Suspension and Entrapped Cell Cultures: Biochemical and Flow Cytometric Analysis as a Function of Serum Content," J. Biotechnology, in press. Dean, R. D. Jr., (1989) 'U.S. Bioprocess Equipment Manufacturers Must Improve Products to Stay Competitive,' Genetic Engineering News, 9, No.8, 6-7. Murakami, H., (1989) 'Serum-free Media Used for Cultivation of Hybridomas', in Monoclonal Antibodies: Production and Application, A. Mizrahi, Ed., Alan R. Liss, Inc., New York, pp. 108-133. Wolpe, S. D., (1984) 'In Vitro Immunization and Growth of Hybridomas in serum-free Medium", In Mammalian Cell Culture, Ed. J. P. Mather, Premium Press, New York.
8.
Kovar, J., (1989) "Various Cell Lines Grow in Protein-free Hybridoma Medium," In Cellular & Developmental Biology, 25, 395-396.
9.
Briand, P., and Lykkesfeldt, A. E., (1984), "Effect of Estrogen and Antiestrogen on Human Breast Cancer Cell Line MCF-7 Adapted to Growth at Low Serum Concentration," Cancer Research, 44, 1114-1119.
10. Ozturk, S. S., and Palsson, B. O. (1989a) "Physiological Changes During the of Hybridoma Cells to Low Serum and Serum-free Media," Submitted. 11.
vitro
Adaptation
Briles, D. E., Forman, C, Hudak, S., and Claflin J. L. (1984) "The Effects of Idiotype on Ability of IgGi Antiphosphorylcholine Antibodies to Protect Mice From fatal Infection with Sfrepfococcus pnetimoniae," Eur. J. Immunol., 14, 1027-1030. the
12.
Dalili, M. and Ollis, (1989) "Transient Kinetics of Hybridoma Growth and Monoclonal Antibody Production in Serum-Limited Cultures", Biotech. Bioeng., 33, 984-990.
13.
Ozturk, S. S., and Palsson, B. O. (1989b) "Effect of Initial Cell Density on Hybridoma Growth, Metabolism, and Monoclonal Antibody Production", J. Biotechology, accepted subject to change. Address reprint request to: Bernhard O. Palsson Department of Chemical Engineering Herbert H. Dow Building University of Michigan Ann
Received for publication November 2, 1989 Accepted after revision January 15, 1990 175
Arbor, MI 48109
Loss of Antibody Productivity During Long-Term
Cultivation of a Hybridoma Cell Line in Low Serum and Serum-Free Media SADETTIN S. OZTURK and BERNHARD
0.
Department of Chemical Engineering, University of Michigan,
PALSSON
Ann Arbor, MI
ABSTRACT A murine hybridoma cell line (167.4G5.3) was adapted to grow in different serum concentrations over a six month period of time. Adaptation to low serum and to serum-free media improved growth rates, but at'low serum (1.25%) the antibody productivity was diminished. Flow cytometric analysis showed the presence of two distinct cell populations with respect to intracellular and surface antibody concentrations. The loss in antibody productivity during adaptation could be attributed to the appearance of a low antibody-containing cell population. Cultures maintained at high serum concentrations did not loose the original high antibody productivity. In a separate experiment the kinetics of growth improvement and loss of antibody production were studied for adaptation from 5% to 1.25% serum-containing media. Over a time period of about four months, the population shifted completely from high-producing cells to low-producing ones in response to the 1.25% environment. A shift-up from 1.25% to 20% serum resulted in the elimination of the low producing population. These results suggest that, for the cell line used, serum-containing factors prevent the loss of antibody productivity.
INTRODUCTION
Hybridoma cell technology is widely used for the preparation of monoclonal antibodies (MAbs) wide spectrum of antigens. The method of establishing permanent cell lines capable of producing antibodies directed to predefined immunogen is based on the fusion of immune lymphocytes with myeloma cells adapted for growth in tissue culture conditions (Esshar, 1985). Several genetic processes occur during and after cell fusion. During cell fusion, the genes from lymphocyte and myeloma fusion partners are rearranged to generate the set of chromosomes that are retained in the resulting hybridoma cell. The genes may continue to be rearranged after the fusion, and chromosomal losses have been reported over longer periods of time (Hunkeler ei al, 1989). The appearance of a low producing hybridoma population in long-term cultures has been observed experimentally (Frame ei al, 1989; Heath et al, 1989). This genetic instability is very important to the economics of long-term hybridoma cell culture (Dean, 1989). Frame et al (1989) showed a decrease with time in the overall antibody production rate that was calculated from the to
a
antibody concentrations and the total viable cell counts in a chemostat. Their flow cytometric analysis revealed heterogeneity of the cell population: a fraction of the cells were lacking the antibody (non-producing cells), while the rest of the cells were producing at a constant specific (per cell) production rate (producing cells). The loss in the observed antibody production rate 167
attributed to the increase in the size of non-producing cell population. A similar observation reported by Heath et al (1989). The non-producing cell population size increased in time and the specific antibody production rate decreased correspondingly. Cell culture media typically contain serum to support cell growth. Although growth is greatly enhanced in the presence of serum, economical and processing considerations favor the use of low serum, or serum-free media for large-scale cultivation. The cost of serum, lot to lot variations in serum quality, undefined composition, and the complications in antibody purification that results from the presence of serum protein have stimulated research on the development of serum-free media. Several serum-free formulations have been developed and are used successfully for a variety of mammalian cells. Antibody productivities in serum-free media were reported to be comparable to those obtained in serum-containing media (Murakami, 1989). Adaptation is needed to obtain improved cell growth in low serum or serum-free media (Wolpe, 1984). Cells have to be passed for several generations in the low, or zero, serum-containing media to achieve good cell viability. The number of generations needed for adaption depends on the cell type and the media used; some cells adapt after a few generations, some cells never adapt (Kovar, 1989). Success of the adaptation of cells to lower serum-containing media can be improved in a 'weaning' procedure in which the serum concentration in the passing media is decreased gradually was was
(Wolpe, 1984).
Several processes may take place during adaptation. Improvement in the growth rate is one of the changes that occur with adaptation to low serum-containing media. The improvement in growth rate suggests changes in the cells' response to any growth-promoting factors provided by serum. For instance, it has been shown that for human breast cancer cell line MCF-7 the number of estrogen receptors vary by the adaptation process (Briand and Lykkesfeldt, 1984). Elsewhere, we have reported that adaptation of hybridoma cells to low serum did not alter key metabolic rates and yield parameters (Ozturk and Palsson, 1989a). Although hybridoma metabolism is insensitive to adaptation to low serum levels, adaptation may alter the specific antibody productivity of hybridoma cells. There is very little information in the literature in this respect. In the present paper, we focus on changes in the antibody production rate that occur during adaptation and present a study of the kinetics of the loss of specific antibody productivity.
MATERIALS and METHODS
line, medium and culture maintenance The murine hybridoma cell line, 167.4G5.3, was kindly provided by Dr. J. Latham Claflin from the University of Michigan Medical Center. The antibody produced by this cell line is an IgGi, directed against the phosphorylcholine (Briles ei al, 1984). Hybridoma cells were made by fusion of BALB/c spleen cells with the nonsecreting plasmacytoma fusion line P3X63-Ag8.653. Antibody was generated from mice immunized with PC-keyhole limpet henemocyanin (KLH). The cells were maintained in 75 cm2 plastic T-flasks (Bélico Glass, Inc., Vineland, NJ) in Iscove's Modified Dulbecco's Medium (IMDM, Gibco Laboratories, Grand Island, NY) containing Fetal Bovine Serum (FBS, Gibco) supplemented with 100 Units/ml potassium penicillin G, and 100 /Jg/ml streptomycin sulphate (Sigma Chemical, St. Louis, MO). The cells were kept at 37° C and 5% COi atmosphere in humidified incubators (VWR Scientific, San Fransisco, CA). Cell
\
Adaptation of cells to low serum levels Cells were adapted to grow in different serum
concentrations by a stepwise procedure that lasted for six months. The cells were initially growing in IMDM with 20% serum (which we designate as 20%-cells). A sample of these cells were frozen and the rest were adapted in 10% serum by passing the cells in IMDM containing 10% serum (10%-cells). After two months of passage, the cells were passed in 5% serum (designated as 5%-cells). Similarly cells were adapated in 2.5% (designated as 2.5%-cells), in 1.25% serum (designated as 1.25%-cells), and in serum-free media (OPTI-MEM, Gibco). Cells were also adapted in one step to grow in 5% and 1.25% serum: exponentially growing 20%-cells were inoculated directly to 5% and 1.25% serum-containing IMDM. The success of this direct adaptation was dependent on the viability of the 20%-cells. Also, the removal of spent media increased the success of the adaptation. Similarly, cells were adapted to 1.25% serum from 5%. In this case, the course of adaptation was followed very closely. Cells were passed every two days and counted using a hemacytometer. The cells were then passed by a dilution ratio determined from the viable cell count. In this procedure, the cell concentration after each passage was kept at 4 104 viable cells/ml. The cell supernatants were saved for the analysis of metabolite and antibody concentrations. -
168
Effect of adaptation
on
cell
physiology
The frozen adapted cells were thawed and propagated in the incubator using T-flasks for batch kinetic study of growth, metabolism, and antibody production rate. Variable passing ratios were applied in order to get all the cells to high concentrations and into exponential phase of growth. The changes in cell behavior were examined in different serum concentrations. Each adapted %-cell was inoculated into IMDM containing different serum concentrations. Four different serum levels were used for this purpose: 1.25, 2.5, 5, and 10%. Each adapted %-cell was centrifuged at 1000 rpm (200 g) for 10 min to remove the spent medium. The cells were washed with fresh media with IMDM containing 1.25% serum and finally, they were distributed into spinner flasks containing different serum concentrations. Experiments were run in duplicates in 100 cm3 spinner flasks (Bélico Glass, Inc., Vineland, NJ) with a 50 cm3 working volume. The flasks were placed on magnetic stirrers (Bélico) in a humidified CO* incubator at 37°C with 5% COi. The agitation rate for the spinner flaks was at 100 rpm. A one milliliter sample was taken twice daily during exponential growth, but once per day during the death phase. After performing cell counts, the samples were centrifuged and the supernatants were stored at -80"C for subsequent determination of metabolite and monoclonal antibody concentrations. Antibody, IgGi, was quantified using an enzyme linked immunosorbent assay (ELISA) as described by Ozturk and Palsson (1989b). a
Flow
Cytometry
A Coulter EPICS 751 flow cytometer was used to analyze the antibody content of the cells. Forward angle light scatter (FALS) and 90" light scatter (90LS) data was used for the estimation of cell size and cell complexity. Viable cells could be differentiated from the dead cells by the differences in their FALS and 90LS characteristics. Non-viable cells were eliminated by gating. Cellular composition of the cells was determined by using fluorescent dyes. Propidium Iodide (PI) was used for the measurement of DNA and RNA content of the cells. Stock solutions of propidium iodide were prepared from powder (Sigma) as 1 mg/ml level using phosphate buffer saline (PBS) and stored in 1 ml aliquots at -80°C Cells were treated with RNase (Sigma) to cleave the RNA for the measurement of DNA content and DNA distribution of the cells. Likewise, DNase (Sigma) was used for the RNA analysis. Stock solutions of DNase or RNase were prepared in PBS as 1 mg/ml and stored in 1 ml aliquots at -80°C. Intracellular protein was stained with Fluorescein Isothiocynate (FITC, Sigma). The stock solutions of FITC were prepared as 1 mg/ml and stored in 1 ml aliquots at -80°C. Cells were stained for the surface and total intracellular IgGi by Fluorescein Isothiocynate conjugated goat anti-mouse IgGi (GAM-FITC, Southern Biotech). The stock concentration of GAM-FITC was 0.02 mg/ml. Cells were fixed with ethanol for analysis of DNA, RNA, intracellular IgGi content. Only, surface IgGi determination was carried out with intact cells. The cells were fixated as follows: cell samples were spun at 1000 rpm for 5 min and the cell culture media was removed. Ceils were then washed with PBS at pH 7.4 and spun down again for 5 min. The cell pellet was resuspended in 70% ethanol and the suspension was kept at A°C overnight. For staining, these fixed cells were spun down at 1000 rpm for 5 min and washed with PBS. The cells were then centrifuged for 5 min and the pellet was suspended in 200 fil GAM-FITC solution. The cells were incubated for 45 min at 4°C. Then the cells were washed with PBS and stained with a mixture of PI and RNase (for DNA determination) or DNase (for RNA determination). The stock solutions of PI and RNase (or DNase) were diluted in PBS to give a PI concentration of 20 fig /ml and a RNase (or DNase) concentration of 40 pg/ml. This double staining was carried out at 37"C for 30 min. The cells were finally washed with PBS and run in the flow cytometer. Protein content was measured separately with fixed cells. The protocol was the same for the measurement of intracellular IgG decribed above, with the exception that FITC was used instead of GAM-FITC. The surface antibody was measured with intact cells (no fixation), as mentioned above. The cell samples were washed with the PBS and stained directly with GAM-FITC for 45 min at i°C. The cells were then washed with PBS and run in the flow cytometer. RESULTS
Effect of adaptation
on
the cell
growth
and monoclonal antibody
production
The short-term response of adapted cells was assessed by inoculating them into batches containing different serum concentration and they were grown in batch culture for a period of two weeks. The growth rates were evaluated in the exponential growth phase and are presented in Figure 1A as a function of serum concentration (short-term) used in these batches. For all
169
the
rate increased by serum concentration. Qualitatively, the concentration was the same for all the adapted cells and was in agreement with the data of Dalili and Ollis (1989) and of Ozturk and Palsson (1989b): the effect of serum on the growth rate was pronounced at low concentration but a constant growth rate was reached at high serum levels. However, quantitatively there was a difference in the responses of adapted cells. Although all the cells originated from the same lot (20%-cells), after adaptation, they resulted in different growth rates for a given short-term serum concentration. All the cells grew slowly at low serum concentrations, i.e. 1.25% FBS. At this serum concentration, the 1.25%-cells showed the highest growth rate. The data shows that the cells acquire growth advantage by adaptation to lower serum concentrations. At any serum concentration, the 1.25%-cells and 2.5% cells grew most rapidly, followed by 5%-cells, and so forth. The growth rate of 1.25%-cells was almost twice that of the original 20%-cells in 1.25% serum. We have seen similar growth rate improvements in another hybridoma cell line by adaptation (Ozturk and Palsson, 1989a). Similar studies have been reported for human breast cancer cell line MCF-7 (Briand and Lykkesfeldt, 1984), and for various hybridoma cells (Kovar, 1989).
adapted cells, the specific growth
response of cells to
new serum
(B)
u II S.Ï i
¡3 Adaptation
Écrum,
%
S
1*
Adaptation
IS
2»
25
«mm, *t/t
Figure 1. Response of adapted cdla to différent aerum concentration» after adaptation. (A) apecific growth rates as a function of aerum concentration uaed in batch growth for different adaptations, (B) apecific antibody production ratea as a function of adaptation aerum levels in different aerum concentrations uaed in batch growth. The antibody production rates for the adapted cells are presented in Figure IB as a function of concentration used for adaptation. The specific antibody production rates were evaluated at different serum concentration for the adapted cells in the batch growth experiment. This plot summarizes both long-term (adaptation) and short-term effect of serum on the antibody productivity of the cells. The specific production rate was not influenced by the short-term exposure to serum levels (response to different serum levels indicated by different symbols in Figure IB). However, there is clearly an influence of adaptation on antibody productivity. The 5%-, 10%-, and 20%-cells showed essentially the same specific antibody productivity, whereas significant losses in productivity were observed for cells adapted to 1.25% and 2.5% serum. These cells did not grow in 1MDM without serum. However, they did grow in OPTI-MEM serum-free media. Although OPTI-MEM is not a serum-free reference point for serum-containing IMDM media, it is instructive to see that the trends observed with reducing serum are continued to serum-free OPTI-MEM. The same loss in antibody productivity was observed when the cells were adapted to OPTI-MEM serum-free media. The gross macromolecular composition of the cells was measured. Figure 2 shows the data on cell size (FALS), DNA, RNA, gross macromolecular protein and intracelluler IgGi content. The data is presented as histograms or distribution functions in which the channel number corresponds to the intensity of the property considered and frequency corresponds to the cell number. The cells were in the exponential growth phase at the time of analysis. The cells were almost identical in terms of size, DNA, RNA, and protein content (Figure 2). There was, however, a significant difference in the antibody content. The 5%-, 10%- and 20%-cells have essentially identical intracellular IgGi levels. The 1.25%- and OPTI-MEM-cells, on the other hand, had a second lower peak along with the same peak exhibited by the cells adapted to higher serum concentrations.
serum
170
Size
DNA
RNA
protein
log(IgG)
OPTI-MEM CmMmei namber
Figure 2. Flow cytometric data on tbe cellular composition of the adapted ceDa. Cell aise, DNA, RNA, and protein content of the cells were almost identical. However, the intracelhilar IgGi content was different in the celia adapted in 1.25%- and in OPTI-MEM-cella.
The appearance of two peaks is further demonstrated in Figure 3. In this two-parameter diagram, the cells are compared with respect to the antibody content and DNA structure. The original cells shown in panel A exhibit a single population. The cells have antibody in all stages of the cell cycle. The adapted cells (1.25%-cells) have the same population and a similar, but lower, antibody-containing population. This second population of cells was very similar in terms of IgGi distribution during the cell cycle and in their cell cycle behavior. contour
i
i
y
12 1Ü
ADAPTED (L23ft FBS)
ORIGINAL (20% FBS)
e-
:i
as
S
! 3d
l
4 3
t >•
is
....Q^£6seSfei >r
...o...%
.< >.
o-
o
..s*.
IT
a....«.....«...*..
»s.
ras-1
1.25*
i
1»
-
/.!• •«
i:iy»~ 1.«
4*
St
6»
!••
121
I
2«
A. Cell
growth during adaptation
rate can be
evaluated from:
where
is the
to
1.25% ~
serum, and B.
Figure
7. The intracellular
in cell number observed
antibody
content
corresponding antibody lévela.
The
growth
ln(i/To) 2
_
x/x0
»•
Time, br
11
expansion
«t
4*
Time, days
•
• •
1J
Figure 6.
single
^g^^^ ^ .
3.« 2J
a
aemm.
days
over
measured
the
two-day period between aubculturing the cells.
during adaptation
173
to
1.25%
aerum
(Figure 6).
that the initial cell concentration after each passage was constant at 4 104 cells/ml. The cells in 5% serum were also passed under the same protocol as a control. Figure 6 shows the increase in cell count during each two-day passing cycle. The growth rate is related to this ratio by fi ln(i/i0)/2 days. After being introduced to 1.25% serum, the cells doubled every 2 days for about a 20-day period. Then, the growth rate increased, and over a 3-month period, the growth rate reached the growth rate observed in 5% serum. The antibody production rate and the intracellular antibody content was followed using ELISA (Figure 6B) and flow cytometry (Figure 7). The cells in 1.25% serum began losing their antibody level after one month, and the fraction of low producers began to increase. •
—
i
—— day,
Figure 8. (A) The fraction of high producers and (B) to
4*
the overall
antibody production rates during the adaptation
aerum.
DISCUSSION The present study has expanded our knowledge of the kinetics of the loss of antibody productivity in long term hybridoma cell cultures and how it is influenced by serum. When 167.4G5.3 hybridoma cells are adapted over a long period of time to low serum-containing media or serumfree media, the cells lose their antibody productivity. The loss of antibody productivity may be directly linked to the appearance of low-producing cells. For a mixed population of high- and low-producers, the apparent antibody production rate will be a cell-density weighed-average of the specific production rates of the high and low producers. The apparent antibody production rate would be equal to the specific antibody production rate of 5%-, 10%- and 20%-cells only when the fraction of low producers is zero. The production rate of the high producer is expected to be qn = 0.2 pg/cell/hr as it is in the 5%- and other high serum-adapted cells. The production rate of low producers is qr, and if the fraction of the low producers is / then the overall antibody production rate is: q
=
?/f(l-/) + /?£
Then, if the fraction of low producers is high, the observed antibody production will be low. Antibody productivities were calculated using the cell counts and the antibody concentrations given in Figure 6 for cells in 5% and 1.25% serum and are presented in Figure 8B. This overall antibody production rate was constant for the cells kept in 5% serum (control). However, the overall productivity decreased over time in 1.25% serum culture. This decrease was in parallel to the decrease in the fraction of high producers in 1.25% serum determined from Figure 7 and plotted in Figure 8A. Hence as the above equation predicts, we see a decrease in the overall antibody production rates that corresponds to the increasing number of low antibody containing cells in the culture. The loss in antibody productivity is thus directly related to the appearance
of the cells with a low antibody content. There are three possible hypotheses to explain our observations. Firstly, we could be observing the appearance of a second clone in the culture, created by genetic drift, that has lost or has reduced ability to produce antibody. The observations we report thus are, simply, the dynamics of a two-population system where the cell composition of the culture shifts with culture conditions. If this hypothesis is true, our results indicate that higher levels of serum support the growth of a high producing clone, whereas long-term adaptation to lower serum-containing media favors the growth of a low- or non-producing clone. A second hypothesis would relate to regulation of 174
transcription or translation rate. In this case, one would hypothesize that the cells gradually change their surface receptor portfolio with time, and this slow change would then, in turn, alter the stability of mRNA or the expression rate of the antibody-coding genes. A third possibility is that the cells might be producing an antibody molecule that is not IgGi and is thus not detected by our assays. Our present data cannot conclusively determine which mechanism might be at the
work. ACKNOWLEDGEMENTS This work was supported by National Science Foundation Grant No. EET-8712756. The authors thank Dr. J. Latham Claflin for providing the hybridoma cell line used in this study and Mark Cameron for his assistance with flow cytometry.
REFERENCES 1.
2. 3. 4.
5. 6.
7.
Esshar, Z., (1985) "Monoclonal Antibody Strategy and Technique," In: Hybridoma Technology in the Biosciences and Medicine, T. A. Springer, Ed. Plenum Press, New York,
3-41, 1985. Hunkeler, N., Hu, VV-S., and Srienc, F., (1989) "Hybridoma Subpoulations Affecting Culture Productivity," In Proceedings of the Annual American Chemical Society Meeting. Frame, K, Sen, S., and Hu, W-S., (1989), "The loss of Antibody Productivity in Continuous Culture of Hybridoma Cells," Biotech. Bioeng., accepted. Heath, C. A., Dilwith, R., and Beifort, G. (1989) "Methods for Increasing Antibody Production in Suspension and Entrapped Cell Cultures: Biochemical and Flow Cytometric Analysis as a Function of Serum Content," J. Biotechnology, in press. Dean, R. D. Jr., (1989) 'U.S. Bioprocess Equipment Manufacturers Must Improve Products to Stay Competitive,' Genetic Engineering News, 9, No.8, 6-7. Murakami, H., (1989) 'Serum-free Media Used for Cultivation of Hybridomas', in Monoclonal Antibodies: Production and Application, A. Mizrahi, Ed., Alan R. Liss, Inc., New York, pp. 108-133. Wolpe, S. D., (1984) 'In Vitro Immunization and Growth of Hybridomas in serum-free Medium", In Mammalian Cell Culture, Ed. J. P. Mather, Premium Press, New York.
8.
Kovar, J., (1989) "Various Cell Lines Grow in Protein-free Hybridoma Medium," In Cellular & Developmental Biology, 25, 395-396.
9.
Briand, P., and Lykkesfeldt, A. E., (1984), "Effect of Estrogen and Antiestrogen on Human Breast Cancer Cell Line MCF-7 Adapted to Growth at Low Serum Concentration," Cancer Research, 44, 1114-1119.
10. Ozturk, S. S., and Palsson, B. O. (1989a) "Physiological Changes During the of Hybridoma Cells to Low Serum and Serum-free Media," Submitted. 11.
vitro
Adaptation
Briles, D. E., Forman, C, Hudak, S., and Claflin J. L. (1984) "The Effects of Idiotype on Ability of IgGi Antiphosphorylcholine Antibodies to Protect Mice From fatal Infection with Sfrepfococcus pnetimoniae," Eur. J. Immunol., 14, 1027-1030. the
12.
Dalili, M. and Ollis, (1989) "Transient Kinetics of Hybridoma Growth and Monoclonal Antibody Production in Serum-Limited Cultures", Biotech. Bioeng., 33, 984-990.
13.
Ozturk, S. S., and Palsson, B. O. (1989b) "Effect of Initial Cell Density on Hybridoma Growth, Metabolism, and Monoclonal Antibody Production", J. Biotechology, accepted subject to change. Address reprint request to: Bernhard O. Palsson Department of Chemical Engineering Herbert H. Dow Building University of Michigan Ann
Received for publication November 2, 1989 Accepted after revision January 15, 1990 175
Arbor, MI 48109
