79. Validation of Viable Producer Cell Removal/Inactivation from Clinical-Grade Oncoretroviral Vector Supernatant Produced in Cell Factories

Abstract

We have developed a semi-closed system to manufacture 8 liter batches of clinical grade PG13-derived oncoretroviral vector supernatant (VS) using 10-tray Nunclon cell factories. The VS (2.6 L per harvest) is collected on each of three (3) consecutive days. After each harvest, the VS is filtered utilizing a modified version of the step-filtration method developed by Reeves and Cornetta (Gene Ther. (2000) 7: 1993–8). The VS is transferred via peristaltic pump through a graded series of depth filters clinically approved for blood filtration. After filtration, the two first harvests are stored at 4°C and pooled with the 3rd harvest. The VS is subsequently stored at −80°C in 250 mL unit-dose aliquots in Cryocyte Freezing Containers. For large-volume supernatant harvests, the step-filtration system yields superior vector titers as compared to traditional membrane filters, due to elimination of membrane fouling. The vector titers determined on HeLa cells are in the range of 1–2×105 TU/mL and allow for 25% transduced T lymphocytes after one round of transduction. To support its use for clinical studies involving ex vivo cell transduction, we validated the step-filtration system to ensure the absence of viable vector-producing cells from each VS unit-dose. A worst-case scenario would involve all cells from the four (4) Cell Factories (approx. 1010 total cells) detaching during a single 2.6 L harvest, resulting in a viable cell burden of 9.6 × 108 cells per unit dose. Therefore, it would be desirable to ensure a minimum 9 log10 reduction in viable cells. Initially, we developed an assay to detect low numbers of viable PG13 packaging cells spiked into 250 mL cell culture medium. We were able to reproducibly (N = 3) detect as few as 3 residual cells per 250 mL medium in a 14-day assay, by staining cell colonies with carbol fuschin. Using this assay, we then performed three experiments in which 2.6 L medium was spiked with 1.13, 1.39, and 1.13 × 109 viable PG13 cell, respectively, followed by step-filtration as described above. The filtered sample was evaluated for the presence of residual cells using the staining colony assay. In all cases, <3 viable cells per 250 mL (assayed in triplicate) were detected, thereby establishing a minimum cell reduction of 8.57 log for the step-filtration procedure. As this does not quite reach the 9 log desired reduction, we chose to examine the −80°C freezing process as a means of achieving additional viable cell reduction. In this experiment, we spiked 250 mL medium, in triplicate, with varying numbers of PG13 cells (30, 300, 3000, and 30,000 cells), followed by transfer to Cryocyte Freezing Containers and storage at −80°C for seven (7) days. The spiked medium aliquots were then thawed and the presence of viable cells evaluated using the assay described above. In this study, viable cells were detected in 1/3 samples at the 30,000 cell level. No samples contained viable cells at any of the lower three spike levels, thereby providing an additional 3 log viable cell reduction, yielding a total viable cell reduction of 11.57 log. This net log viable cell reduction is more than sufficient to ensure the absence of residual viable vector-producing cells from a unit-dose of clinical-grade retroviral VS.