Serum-free production, concentration and purification of recombinant retroviruses.

Concentrated and purified retroviral preparations are useful in gene therapy protocols and in the study of retroviral function. For the transduction of refractory cell types, such as hematopoietic stem cells and for in vivo applications such as liver-directed gene therapy after induction of liver regeneration, gene transfer efficiencies can be increased by concentrating packaging cell-line supernatants (3,4). The study of retroviral fusion with cellular or model membranes has been hampered by the difficulty in obtaining high titer-purified viral preparations (8). A major problem in retroviral concentration is the instability of the viral particles. During purification procedures, shear forces dissociate the envelope proteins from the gag core of the retroviruses, leading to a noninfectious particle (2,7,12,16). Several methods for the concentration of retroviruses have been described. Retroviruses can be concentrated by passing producer-line supernatant through filtration devices (6,9, 13,14). The main disadvantage of these methods is that most media components are co-concentrated as well. In case of an in vivo application, this may generate a strong immune response in the treated animal. The presence of animal products, such as bovine serum, in the cell culture media is a concern because of the possible presence of viruses or prion proteins. Large amounts of nonviral proteins will disturb assays to monitor membrane fusion. The concentrate will also contain cytokines and growth factors from bovine serum or produced by the packaging cells. These cytokines and growth factors might subsequently interfere with the target cells. Furthermore, the concentrated supernatant will also contain inhibitors of viral function, such as free envelope proteins and proteoglycans (1,10). More elaborate procedures to concentrate and purify retroviruses have been described. These procedures involve multiple steps including density gradient centrifugation, or gel filtration, and are consequently time-consuming and have a low yield (3,4). By combining the production of retroviruses in serum-free media and a simple low-speed centrifugation procedure, we were able to concentrate and purify retroviruses in one step. A retroviral producer line, based on PA317, producing the LNFZ vector (15) capable of transferring the E. coli β-galactosidase cDNA was used for all experiments described in this paper. Producer lines were cultured in Dulbecco’s modified Eagle medium (DMEM; Life Technologies, Gaithersburg, MD, USA), supplemented with 2 mM L-glutamine and 100 U/mL penicillin-G/streptomycin plus 10% fetal calf serum (FCS; Summit Biotechnology, Fort Collins, CO, USA) until confluent. Viral supernatants were collected from 13.5-cm-diameter plates in 15 mL of medium. The cells were washed with D-MEM without serum, and serum-free medium was added. After 24 h, the supernatant was collected, filtered through a 0.22-μm filter and used for concentration or titration. Retrovirus titers were determined by diluting the different retroviral preparations 104–105 and infecting 5 × 105 NIH3T3 TKcells in 6-cm-diameter plates in the presence of 4 μg/mL polybrene overnight. The day after infection, the cells were split 1/10 and plated in DMEM, 10% FCS and 1 mg/mL G-418 (Life Technologies). After one week, the colonies were fixed and stained in 50% methanol/10% glacial acetic acid/ H2O containing 0.1% Coomassie Brilliant Blue G250 and counted. In initial experiments, it was found that culturing of retroviral producer lines in D-MEM without FCS led to a significant decrease in viral titers, as previously published (13). Several serumfree formulations were tested. OPTI-MEM, D-MEM/F-12 and viral production (VP) media were obtained from Life Technologies. HL-1 supplement was obtained from BioWhittaker (Walkersville, MD, USA). OPTI-MEM and VP yielded virus at comparable titers to D-MEM plus FCS. Titers in DMEM/F-12 + HL-1 were slightly lower (Table 1). For further experiments, VP was chosen because this medium does not contain animal products and because it has the lowest protein content of all tested media (5 μg/mL). In disposable 50-mL tubes (Corning Costar, Cambridge, MA, USA), 15 mL of packaging supernatant were collected. The tubes were centrifuged in a 40-cmdiameter swing-out rotor at 3600 rpm for 20 h at 4°C. The maximum g force was 2900×. The supernatant was carefully removed, and the invisible pellet was resuspended in 0.5 mL of D-MEM without serum by pipetting up and down and gently vortex mixing. The titers of the resuspended pellet and the supernatant were determined. The results of 5 independent purification procedures are summarized in Table 2. The average purification factor was 12. Total recovery of virus, supernatant and pellet was 87%, indicating that at these low g forces, only minor damage to the retrovirus occurs.