Abstract

Recombinant VWF was co-expressed with recombinant FVIII in CHO cells. In order to obtain fully processed, mature rVWF, rVWF was exposed to recombinant CHO-cell-derived furin for VWF pro-peptide removal. Fermentation for both rVWF and furin and downstream processing were performed under serum-free conditions.

Mature rVWF was highly purified by a series of conventional chromatography steps to a purity of > 95%. rVWF was formulated in a protein-free buffer.

Several large scale batches of rVWF were produced and characterized in comparison to Humate-P, Wilfactin and a highly purified plasma-derived (pd)VWF, an experimental pd VWF product obtained from cryoprecipitate. Residual pro-VWF levels were approximately 2% of total VWF antigen compared with 0.12, 0.05 and 0.13 for the rVWF product, experimental pd VWF, Humate-P and Wilfactin, respectively. VWF pro-peptide levels were approximately 0.5 pmol per unit VWF antigen, compared with 0.6 and 0.9 for Humate-P and Wilfactin, respectively; in the experimental pd VWF pro-peptide was below the limit of detection. Residual FVIII content was approximately 0.02 units (chromogenic assay) per unit VWF antigen, compared with 0.4 units in Humate-P, and 0.04 units in Wilfactin, and ≤ 0.01 in the experimental pd VWF product.

Ristocetin cofactor activity was around 0.7 VWF:RCo units per unit antigen for all rVWF and pd VWF products tested. Collagen-binding activity was also similar between rVWF and pd VWF controls (around 0.8 – 0.9 units per unit). FVIII-binding capacity of rVWF was compared by setting the mean pd VWF as 100% and was found to also be ± 100%. The VWF affinity for FVIII was similar in rVWF and pd VWF (Biacore). rVWF showed high molecular weight multimers in low resolution agarose gels, with a pre-dominance in the high molecular weight range. In contrast to pd VWF, rVWF showed homogenous multimer bands without triplet structure formation due to the non-exposure to ADAMTS13 during the manufacturing process.

The pharmacokinetic properties of rVWF were compared with pd VWF in VWF knock-out mice. The half-life or rVWF in VWD mice was 4.3 and 3.5 hours for 2 batches investigated, which seems longer than the half-lives obtained for pd VWF, 2.9 hours for Humate-P and 2.2 hours for a highly purified pd VWF control. Both VWF preparations stabilized endogenous murine FVIII which rose over time to the levels observed in normal C57BI/6J mice.

Overall, these results showed that the properties of mature, furin-processed rVWF in vitro and in vivo are equivalent to those of plasma-derived VWF.

Disclosures: All authors are employees of Baxter AG.; Authors have options to Baxter stock.

Author notes

*

Corresponding author