Large scale production of recombinant mouse and rat growth hormone by fed-batch GS-NSO cell cultures

Investigations of biological effects of prolonged elevation of growth hormone in animals such as mice and rats require large amounts of mouse and rat growth hormone (GH) materials. As an alternative to scarce and expensive pituitary derived materials, both mouse and rat GH were expressed in NSO murine myeloma cells transfected with a vector containing the glutamine synthetase (GS) gene and two copies of mouse or rat GH cDNA. For optimal expression, the mouse GH vector also contained sequences for targeting integration by homologous recombination. Fed-batch culture processes for such clones were developed using a serum-free, glutamine-free medium and scaled up to 250 L production scale reactors. Concentrated solutions of proteins, amino acids and glucose were fed periodically to extend cell growth and culture lifetime, which led to an increase in the maximum viable cell concentration to 3.5×10⁹ cells/L and an up to 10 fold increase in final mouse and rat rGH titers in comparison with batch cultures. For successful scale up, similar culture environmental conditions were maintained at different scales, and specific issues in large scale reactors such as balancing oxygen supply and carbon dioxide removal, were addressed. Very similar cell growth and protein productivity were obtained in the fed-batch cultures at different scales and in different production runs. The final mouse and rat rGH titers were approximately 580 and 240 mg/L, respectively. During fed-batch cultures, the cell growth stage transition was accompanied by a change in cellular metabolism. The specific glucose consumption rate decreased significantly after the transition from the growth to stationary stage, while lactate was produced in the exponential growth stage and became consumed in the stationary stage. This was roughly coincident with the beginning of ammonia and glutamate accumulation at the entry of cells into the stationary stage as the result of a reduced glutamine consumption and periodic nutrient additions.     

Repeated-batch cultures of Baby Hamster Kidney cell aggregates in stirred vessels

Natural aggregates of Baby Hamster Kidney cells were grown in stirred vessels operated as repeated-batch cultures during more than 600 hours. Different protocols were applied to passaging different fractions of the initial culture: single cells, large size distributed aggregates and large aggregates. When single cells or aggregates with the same size distribution found in culture are used as inoculum, it is possible to maintain semi-continuous cultures during more than 600 hours while keeping cell growth and viability. These results suggest that aggregate culture in large scale might be feasible, since a small scale culture can easily be used as inoculum for larger vessels without noticeable modification of the aggregate chacteristics. However, when only the large aggregates are used as inoculum, it was shown that much lower cell concentrations are obtained, cell viability in aggregates dropping to less than 60%. Under this selection procedure, aggregates maintain a constant size, larger than under batch experiments, up to approximately 400 hours; after this time, aggregate size increases to almost twice the size expected from batch cultures.     

Experimental collision efficiencies of polymer-flocculated animal cells.

The determination of particle collision kinetics is useful to decouple the effects of process parameters on individual events in flocculation. This paper discusses the effects of flocculation conditions on the collision efficiency of ATCC strain CRL 1606 hybridomas flocculated with poly-L-histidine. Experimental determinations of the collision efficiency of cells in Couette flow are presented over a range of experimental conditions. The collision efficiency correlates with the cell zeta potential to the -2.4 power at high surface coverage, consistent with literature results in latex systems. At low coverage, accounting for the distribution of polymer on the cells corrects for deviation from the high-coverage behavior. Collision is dependent on the hydrodynamic environment as well. At high surface coverage, collision efficiency is weakly dependent on hydrodynamic conditions and follows a dependency on the shear rate and viscosity to the -0.32 power. This is consistent with ionic coagulation theory. At low surface coverage, the collision efficiency is strongly dependent on the viscous fluid forces. The results versus both dose and shear rate over the entire range of surface coverages are consistent with weak intercell bonding. Collision kinetics in the presence of high molecular weight dextrans show steric hindrance to cell collision.     

Fed-batch culture of recombinant NS0 myeloma cells with high monoclonal antibody production

An amplified NS0 cell line transfected with a vector expressing a humanized monoclonal antibody (MAb) against CD-18 and glutamine synthetase (GS) was cultivated in a 1.5 L fed-batch culture using a serum-free, glutamine-free medium. Concentrated solutions of key nutrient components were fed periodically using a simple feeding control strategy. Feeding amounts were adjusted daily based on the integral of viable cell concentration over time (IVC) and assumed constant specific nutrient consumption rates or yields to maintain concentrations of the key nutrient components around their initial levels. On-line oxygen uptake rate (OUR) measurement was used to aid empirically the adjustment of the feeding time points and amounts by inferring time points of nutrient depletion. Through effective nutritional control, both cell growth phase and culture lifetime were prolonged significantly, resulting in a maximal viable cell concentration of 6.6 x 10(9) cells/L and a final IVC of 1.6 x 10(12) cells-h/L at 672 h. The final MAb concentration reached more than 2.7 g/L. In this fed-batch culture, cellular metabolism shifts were repeatedly observed. Accompanying the culture phase transition from the exponential growth to the stationary phase, lactate, which was produced in the exponential growth phase, became consumed. The time point at which this metabolism shift occurred corresponded to that of rapid decrease of OUR, which most likely was caused by nutrient depletion. This transition coincided with the onset of ammonia, glutamate and glutamine accumulation. With removal of the nutrient depletion by increasing the daily nutrient feeding amount, OUR recovered and viable cell concentration increased, while cell metabolism shifted again. Instead of consumption, lactate became produced again. These results suggest close relationships among nutrient depletion, cell metabolism transition, and cell death. (c) 1997 John Wiley & Sons, Inc. Biotechnol Bioeng 55: 783-792, 1997.     

Modifying TIMER to generate a slow-folding DsRed derivative for optimal use in quickly-dividing bacteria

It is now well appreciated that members of pathogenic bacterial populations exhibit heterogeneity in growth rates and metabolic activity, and it is known this can impact the ability to eliminate all members of the bacterial population during antibiotic treatment. It remains unclear which pathways promote slowed bacterial growth within host tissues, primarily because it has been difficult to identify and isolate slow growing bacteria from host tissues for downstream analyses. To overcome this limitation, we have developed a novel variant of TIMER, a slow-folding fluorescent protein, named DsRed₄₂, to identify subsets of slowly dividing bacteria within host tissues. The original TIMER folds too slowly for fluorescence accumulation in quickly replicating bacterial species (Escherichia coli, Yersinia pseudotuberculosis), however DsRed₄₂ accumulates red fluorescence in late stationary phase cultures of E. coli and Y. pseudotuberculosis. We show DsRed₄₂ signal also accumulates during exposure to sources of nitric oxide (NO), suggesting DsRed₄₂ signal detects growth-arrested bacterial cells. In a mouse model of Y. pseudotuberculosis deep tissue infection, DsRed₄₂ signal was detected, and primarily accumulates in bacteria expressing markers of stationary phase growth. There was no significant overlap between DsRed₄₂ signal and NO-exposed subpopulations of bacteria within host tissues, suggesting NO stress was transient, allowing bacteria to recover from this stress and resume replication. This novel DsRed₄₂ variant represents a tool that will enable additional studies of slow-growing subpopulations of bacteria, specifically within bacterial species that quickly divide.     

Evaluation of a microcarrier process for large-scale cultivation of attenuated hepatitis A

Microcarrier culture was investigated for the propagation of attenuated hepatitis A vaccine in the anchorage-dependent human fibroblast cell line, MRC-5. Cells were cultivated at 37°C for one to two weeks, while virus accumulation was performed at 32°C over 21 to 28 days. The major development focus for the microcarrier process was the difference between the cell and virus growth phases. Virus antigen yields, growth kinetics, and cell layer/bead morphology were each examined and compared for both the microcarrier and stationary T-flask cultures. Overall, cell densities of 4–5×10⁶ cells/ml at 5–10 g/l beads were readily attained and could be maintained in the absence of infection at either 37°C or 32°C. Upon virus inoculation, however, substantial cell density decreases were observed as well as 2.5 to 10-fold lower per cell and per unit surface area antigen yields as compared to stationary cultures. The advantages as well as the problems presented by the microcarrier approach will be discussed.     

Hydrodynamic effects on BHK cells grown as suspended natural aggregates

Baby hamster kidney (BHK) cell aggregates grown in stirred vessels with different working volumes and impeller sizes were characterized. Using batch cultures, the range of agitation rates studied (25-100 rpm) led to aggregates with maximum sizes of 150 mum. Necrotic centers were not observed and cell specific productivity was independent of aggregate size. High cell viability was found for both single and adherent cells without an increase in cell death when agitation rate was increased. The increase in agitation rate affected aggregates by reducing their size and increasing their concentration and cell concentration in aggregates, while increasing the fraction of free cells in suspension. The experimental relationship between aggregate size and power dissipation rate per unit of mass was close to -1/4, suggesting a correlation with a critical turbulence microscale; this was independent of vessel scale and impeller geometry over the range investigated. Viscous stresses in the viscous dissipation subrange (below Kolmogoroff eddies) appear to be responsible for aggregate breakage. Under intense agitation BHK cells grown in the absence of microcarriers existed as aggregates without cell damage, whereas cells grown on the surface of microcarriers were largely reduced. This is a clear advantage for scaleup purposes if aggregates are used as a natural immobilization system in stirred vessels. (c) 1995 John Wiley & Sons, Inc.     

Scaleable production of adenoviral vectors by transfection of adherent PER.C6 cells

Recombinant adenoviruses are efficient gene delivery vectors that are being evaluated in many gene therapy and vaccine applications. Methods for rapid production of ca. 10(12)-10(13) virus particles (VPs) are desired to enable rapid initial evaluation of such vectors. For this purpose, a scalable transfection procedure was developed for production of an adenovirus type 5 vector expressing HIV-1 gag gene (MRKAd5gag). Adherent PER.C6 cells were transfected by calcium phosphate coprecipitation of the linearized, 36 kb adenovirus plasmid in disposable culture vessels. Various process variables including precipitate formation time, DNA concentration, and harvest time were investigated to rapidly achieve desired virus yields using an adenovirus plasmid encoding the green fluorescent protein (pAd5gfp). Using an optimized procedure, consistent production of >5 x 10(10) VPs per 1-tray Nunc cell factory (NCF) with a ratio of infectious units to virus particles of >1:10 was obtained for the MRKAd5gag vector. This scaleable process can be used to produce adenoviral vectors using several 1-tray NCFs or a single multiple-tray NCF within 1 month from the time of plasmid construction.     

Cultivation of attenuated hepatitis A virus antigen in a titanium static mixer reactor

The titanium static mixer reactor, demonstrated for a variety of vaccine processes during the late 197s, was investigated for the production of attenuated hepatitis A virus antigen from anchorage-dependent MRC-5 cells. This reactor system used Charles River Biotechnological Services cabinets for monitoring and process control. Cell inoculation protocols, using 6000-10,000 cells/cm(2), resulted on over 95% attachment at both the laboratory and pilot scales. Indirect monitoring techniques using oxygen, glucose, L-serine, and L-glutamine uptake rates were indicative of cell growth prior to virus inoculation as well as environmental and/or nutrient limitations. Seven laboratory-scale (3900 cm(2)) runs and one pilotscale (265,000 cm(2)) run were conducted to investigate refeeding regiments, parallel versus perpendicular element orientation, increased element surface area per unit volume, and scale-up performance. In general, lysate antigen yields achieved were similar to those of parallel T-flasks cultivated under similar conditions. (c) 1994 John Wiley & Sons, Inc.     

Changes in animal cell natural aggregates in suspended batch cultures

Some anchorage-dependent animal cells can form natural aggregates in stirred tanks. Baby hamster kidney (BHK) natural aggregates are described and characterized. Total cell concentration and viability could be obtained after aggregate mechanical aissociation, with negligible cell lysis and no change in cell membrane permeability. During a normal batch run, aggregates were formed immediately after inoculation, a few spherical aggregates increasing size during the initial growth phase. At the end of the growth phase, an increase in aggregate concentration was observed, without a considerable increase in aggregate diameter. At the end of the batch run, 160 h after inoculation, aggregates disintegrated into smaller, non-spherical units, following a sharp viability decrease. Cell concentrations of 1. 2 · 10⁶ cells/ml were obtained, with 60% of the total cells being in aggregates; the cell concentration in aggregates achieved 5 · 10⁸ cells/ml, with a porosity of 55%. Viability was consistently in the range 85–90%, both for aggregate and suspended cells.