Improved, human-based packaging cell lines allow the production of high-titer, RCR-free retroviral vectors. The utility of these cell lines for the production of clinical grade vectors critically depends on the definition of optimal conditions for scaled-up cultures. In this work, a clone derived from the TE Fly GALV packaging cell (Duisit et al. Hum. Gene Ther. 1999, 10, 189) that produces high titers of a lacZ containing retroviral vector with a Gibbon Ape Leukemia Virus envelope glycoprotein was used. This clone can produce (2-5) x 10(6) PFU cm(-3) in small scale cultures and has been evaluated for growth and vector production in different reactor systems. The performances of fixed bed reactors [CellCube (Costar) and Celligen (New Brunswick)] and stirred tank reactors [microcarriers and clump cultures] were compared. The cells showed a higher apparent growth rate in the fixed bed reactor systems than in the suspension systems, probably as a result of the fact that aggregation and/or formation of clumps led to a reduced viability and reduced growth of cells in the interior of the clumps. As a consequence, the final cell density and number were in average 3- to 7-fold higher in the fixed bed systems in comparison to the suspension culture systems. The average titers obtained ranged from 0.5 to 2.1 x 10(7) PFU cm(-3) for the fixed bed and microcarrier systems, while the clump cultures produced only (2-5) x 10(5) PFU cm(-3). The differences in titers reflect cell densities as well as specific viral vector production rates, with the immobilization and microcarrier systems exhibiting an at least 10-fold higher production rate in comparison to the clump cultures. A partial optimization of the culture conditions in the Celligen fixed bed reactor, consisting of a 9-fold reduction of the seeding cell density, led to a 5-fold increased vector production rate accompanied by an average titer of 3 x 10(7) PFU cm(-3) (maximum titer (4-5) x 10(7) PFU cm(-3)) in the fixed bed reactor. The performance evaluation results using mathematical models indicated that the fixed bed bioreactor has a higher potential for retroviral vector production because of both the higher reactor productivity and the lower sensitivity of productivity in relation to the changes in final retrovirus titer in the range of 3 x 10(6) to 15 x 10(6) PFU cm(-3).

No abstract

The present work studies the physico-chemical properties of retroviral vector membrane, in order to provide some explanation for the inactivation kinetics of these vectors and to devise new ways of improving transduction efficiency. For this purpose, vectors with an amphotropic envelope produced by TE Fly A7 cells at two culture temperatures (37 and 32 6C) were characterized by different techniques. Electron paramagnetic resonance (EPR) results showed that vectors produced at 32 6C are more rigid than those produced at 37 6C. Further characterization of vector membrane composition allowed us to conclude that the vector inactivation rate increases with elevated cholesterol to phospholipid ratio. Differential scanning calorimetry (DSC) showed that production temperature also affects the conformation of the membrane proteins. Transduction studies using HCT116 cells and tri-dimensional organ cultures of mouse skin showed that vectors produced at 37 6C have higher stability and thus higher transduction efficiency in gene therapy relevant cells as compared with vectors produced at 32 6C. Overall, vectors produced at 37 6C show an increased stability at temperatures below 4 6C. Since vector membrane physico-chemical properties are affected in response to changes in culture temperature, such changes, along with alterations in medium composition, can be used prospectively to improve the stability and the transduction efficiency of retroviral vectors for therapeutic purposes.

An electronic nose (EN) device was used to detect microbial and viral contaminations in a variety of animal cell culture systems. The emission of volatile components from the cultures accumulated in the bioreactor headspace, was sampled and subsequently analysed by the EN device. The EN, which was equipped with an array of 17 chemical gas sensors of varying selectivity towards the sampled volatile molecules, generated response patterns of up to 85 computed signals. Each 15 or 20 min a new gas sample was taken generating a new response pattern. A software evaluation tool visualised the data mainly by using principal component analysis. The EN was first used to detect microbial contaminations in a Chinese hamster ovary (CHO) cell line producing a recombinant human macrophage colony stimulating factor (rhM-CSF). The CHO cell culture was contaminated by Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus and Candida utilis which all were detected. The response patterns from the CHO cell culture were compared with monoculture references of the microorganisms. Second, contaminations were studied in an Sf-9 insect cell culture producing another recombinant protein (VP2 protein). Contaminants were detected from E. coli, a filamentous fungus and a baculovirus. Third, contamination of a human cell line, HEK-293, infected with E. coli exhibited comparable results. Fourth, bacterial contaminations could also be detected in cultures of a MLV vector producer cell line. Based on the overall experiences in this study it is concluded that the EN method has in a number of cases the potential to be developed into a useful on-line contamination alarm in order to support safety and economical operation for industrial cultivation.