JANUARY/FEBRUARY 199 1
VOLUME 7, NUMBER 1
0 Copyright 1991 by the American Chemical Society and the American Institute of Chemical Engineers
Influence of Medium Exchange Schedules on Metabolic, Growth, and GM-CSF Secretion Rates of Genetically Engineered NIH-3T3 Cells Jerry Caldwell,t,s Betty Locey,S Michael F. Clarke,i*§Stephen G. Emerson,Ss§Jand Bernhard 0. Palsson**t Departments of Chemical Engineering, Internal Medicine, and Pediatrics and Program in Cell and Molecular Biology, University of Michigan, Ann Arbor, Michigan 48109
T h e metabolic and secretory characteristics of NIH-3T3 fibroblasts transfected with a cDNA encoding human granulocyte-macrophage colony stimulating factor (GMCSF) were examined as a function of the culture medium exchange schedule. T h e rates of glucose a n d glutamine consumption and of lactate and ammonia production were measured over exchange schedules ranging from complete medium replacement weekly (l/week) t o complete medium replacement daily (7/week). All measured metabolic rates increased with increased medium exchange rates and accelerated sharply between exchange rates of 3.5/week a n d 7/week. T h e lactate/glucose and ammonia/glutamine yield coefficients, however, remained invariant a t about 1.9 and 1.0mol/mol, respectively, under all medium perfusion conditions. A shift-up in medium perfusion rates from 3.5/week to 7/week resulted in increased metabolic rates t h a t resembled those observed in the cultures t h a t were exchanged a t the 7/week rate throughout, showing t h a t the metabolic rates could be directly controlled by the perfusion rate. Differential regulation of medium versus serum perfusion demonstrated t h a t increased NIH-3T3 cell metabolism was directly proportional t o the serum flux t o which the cells were exposed. T h u s a limiting serum component is responsible for the altered metabolic and growth rates. T h e GM-CSF production by the transfected 3T3 cells was stable but exhibited substantial transient increases during periods of cell proliferation, demonstrating t h a t the secretion of transfected gene products can be highly modulated even when the cDNA is driven from a constitutive promoter. These studies show t h a t the metabolic and secretory behavior of genetically engineered cells is influenced by the medium exchange schedule.
1. Introduction The defined culture of genetically engineered cells is playing an increasing role in advances in molecular and cellular bioengineering. Transfection of functional cDNAs on plasmids linked to either constitutive or inducible promoters allows the local secretion of specified proteins. Genetically engineered secretory cells can serve as powerful reagents to continuously produce hormones, growth fact
Department of Chemical Engineering.
5
Program in Cell and Molecular Biology. Department of Pediatrics.
* Department of Internal Medicine. 11
tors, and cytokines for highly efficient and specialized culture systems. Once the host cells are stably transfected with the desired expression vector, however, protein secretion and metabolism will be regulated by the specific conditions of the cell culture, including cell concentration, perfusion conditions, and geometric constraints of the culture. While there is rapidly accumulating a great deal of information concerning the molecular genetics of gene expression per se, there is as yet little information on the physiologic behavior of transfected host cells as a function of the key variables of bioreactor design. The use of genetically engineered cells is likely to play
8756-7938/91 /3007-0001$02.50/0 0 199 1 American Chemical Society and American Institute of Chemical Engineers
n
Biotechnol. frog., 1991, Vol. 7, No. 1
I
an important role, in particular, in our attempts to reconstitute tissues ex vivo. For example, the provision of key hematopoietic growth factors via genetically engineered candidate stromal cells could provide essential ingredients to the construction of high-efficiency human bone marrow tissue culture ex vivo (Emerson et al., 1991). Suitable choices of stromal cells, hematopoietic growth factor genes, and specific tissue culture perfusion parameters could theoretically be made to maximize cell production, culture longevity, and lineage specificity. The first step in such targeted tissue engineering approaches, however, requires a more detailed understanding of the metabolic and functional behavior of such transfected candidate stromal cells. We have therefore examined the behavior of genetically engineered NIH-3T3 murine fibroblasts (Jainchill et al., 1969),by measuring fundamental determinants of cellular metabolism and functional secretory activity. In this model, NIH-3T3 cells were transfected with a full-length cDNA encoding the human granulocyte-macrophage colony stimulating factor (GM-CSF) gene regulated by the metallothionein promoter and cytomegalovirus (CMV) enhancer. GM-CSF, a glycoprotein with an apparent molecular mass of 18-23 kDa, stimulates species specific production of both granulocytes and macrophages as well as eosinophil and erythroid progenitor cells in cultures of human or murine bone marrow cells [see Metcalf (1988)]. Our genetically engineered plasmid, when transfected into 3T3 cells, permits the constitutive expression and secretion of GM-CSF at theoretically constant levels. In this study, we report measurements of glycolytic flux, oxidative metabolism, and GM-CSF secretion under variations in medium and serum perfusion rates. The range of exchange schedules used was determined by the two extremes, ranging from typical cell culture protocols that specify a weekly exchange of medium to those estimated to prevail in vivo, which are on the order of a complete exchange of medium daily. The results indicate that NIH-3T3 cellular metabolism is sensitive to perfusion conditions, and the secretion of even a constitutively regulated gene product is directly regulated by the metabolic adaptations to variations in perfusion.
2. M a t e r i a l s a n d M e t h o d s Construction of t h e GM-CSF Expression Vector. The mouse metallothionein promoter (Glanville et al., 1981; Foster et al., 1988; Mueller et al., 1988; Stuart et al., 1986) was digested with EcoRI and BglII, and the approximately 2-kb fragment containing the promoter was inserted into the EcoRI-BamHI fragment of pSP65 (Melton et al., 1984). This plasmid is termed p65MT. Complementary DNA for human GM-CSF (Wong et al., 1985) was then cloned into an EcoRI-PstI fragment of pSP65. This pSP65GM-CSF plasmid was then digested with EcoRI and the overhangwas filled in with Klenow fragment of DNA polymerase I, followed by digestion of the linearized DNA with HindIII. This fragment, which comprised the GM-CSF coding region, was then subcloned into the SalI-filled/HindIII site of p65MT. The 2.7-kb fragment comprising the metallothionein promoter and the GM-CSF coding region was then isolated and placed into pSV2neo (Southern and Berg, 1982) from which the simian virus (SV-40) promoter was removed, resulting in placing the SV-40 polyadenylation signal downstream of the GM-CSF coding sequence. The neomycin (G418) resistance gene was isolated from pSV2neo by digestion of the 3-kb PcuII-EcoRI fragment and addition of EcoRI linkers onto the PvuII site. The resultant neo resistance gene with EcoRI ends was then subcloned into the EcoRI
CY
Y
i
'1 0
0
0
5
10
15
20
Time, (day)
04 0
5
10
15
0
Time, (day)
F i g u r e 1. Glucose concentration and consumption in confluent NIH/3T3 cultures. Medium in duplicate 10-mL cultures was exchanged on each of the following schedules over 18 days in culture: daily complete 10-mL exchange (7/week, O),daily5-mL medium exchange (3.5/week, w), biweekly complete 10-mL medium exchange (2/week, n), and biweekly 5-mL exchange (l/week, 0 ) . An aliquot of the spent medium a t each exchange was assayed for glucose concentration. Glucose consumption was determined by subtracting the glucose concentration in spent medium from the glucose concentration in fresh medium. (A) Glucose concentration in mM day-' flask-'. (B) Glucose concentration in mM day-' flask-'. Table I. Mean Metabolic Rates as a Function of Exchange Schedule8 lactate glutamine ammonia exchange glucose scheduleb consumption production consumption production 7 10.6 19.3 1.4 1.5 3.5 4.3 6.5 0.7 0.6 6.1 0.6 0.6 2 3.4 4.4 0.3 0.3 1 2.9 The rates shown are metabolic consumption/production rates calculated from days 7-18 in units of mM day-' flask-'. The exchange schedule given is equivalent to the number of complete medium exchanges per week: 7 = daily complete medium exchange;3.5 = daily half medium exchange; 2 = biweekly complete medium exchange; 1 = biweekly half medium exchange.
site of the GM-CSF expression plasmid to create the plasmid MTGM-CSFneo. Transfection a n d Selection of GM-CSF/3T3 Cells. The plasmid MTGM-CSFneo was transfected by electroporation of linearized DNA into the NIH-3T3 murine fibroblast cell line. Transformants were selected in medium containing 500 kg/mL G418. Individual single cell clonses were isolated and screened for production of GM-CSF by their simulation of proliferation of the leukemic cell line AML-193, as previously described (Adams et al., 1989). A single clone, whose production of GMCSF was stable over several months in continuous culture, was used for all of the studies that followed. E s t a b l i s h m e n t a n d M a i n t e n a n c e of A d h e r e n t Layers. Adherent layers were established by inoculating 5 x lo3 cells/mL into 75-cm2tissue culture flasks in Iscoves modified Dulbecco's medium (IMDM) supple-
Biotechnol. Prog., 1991, Vol. 7, No. 1
".
0
5
10
3
15
20
15
20
Time, (day)
04 0
I 5
10
Tlme, (day)
F i g u r e 2. Lactate concentration and production in confluent
NIH/3T3 cultures. Cultures were maintained and assayed as described for Figure 1. (A) Lactate concentrations in mM day-' flask-'. (B) Lactate production in mM day-' flask-'.
mented with 1070 fetal calf serum (FCS) (Hyclone) and 1YO penicillin/streptomycin (Gibco). During the subsequent establishment of the adherent layers, medium was exchanged twice weekly. Experiments in which exchange schedules were varied began once the cultures reached visual confluence (approximately 1.1X lo5cells/cm2).Prior to use in this study, the NIH-3T3 cells were maintained on a medium exchange schedule of 50% biweekly (l/week) and were routinely passaged prior to confluence a t a ratio of 1:20. No spontaneous transformed foci were observed in any cultures. Medium Exchange Rates a n d Monitoring of t h e Cultures. Parallel cultures of passaged adherent fibroblasts were exchanged on each of four schedules: (1)half of the medium exchanged twice weekly (1weekly exchange equivalent), (2) all of the medium exchanged twice weekly (2 exchanges weekly), (3) half the medium exchanged daily (3.5 exchanges weekly), and (4) all of the medium exchanged daily (7 exchanges weekly). Triplicate cultures were maintained with each medium exchange schedule, and the data presented are the average of the three cultures. In some experiments, the exchange schedules were shifted between medium exchange rates (assay values from cultures that were shifted represent the average of duplicate cultures). Aliquots of each culture supernatant were assayed for glucose, lactate, glutamine, ammonia, and GM-CSF concentrations. Metabolic Measurements. Glucose and Lactate. Aliquots of removed supernatants were diluted 1:3 with deionized water, and then glucose and lactate concentrations were measured by using a YSI Model 2000 glucose/ lactate analyzer (Yellow Springs Instrument Co). Sample dilution was necessary to reduce the concentration to within the ranges of greatest sensitivities of the analyzer (
