Continuous Capture of Recombinant Adeno-Associated Virus Particles Using Twin-Column CaptureSMB

Recombinant adeno-associated viruses (rAAVs) have gained increasing importance in gene therapy due to their safe and precise gene delivery. However, certain indications require substantially higher vector doses, pushing manufacturing capacity and cost of goods (COG) to its limits. In this study, we present a proof-of-concept evaluation of a continuous twin-column capture process (CaptureSMB) enabling direct purification of rAAV5 from minimally processed, nonlysed perfusion harvest without prior enzymatic nuclease digestion, concentration or buffer exchange. This approach differs fundamentally from conventional batch workflows which typically mandate clarification and concentration before affinity capture and offers a novel process integration in viral vector manufacturing. A single-column batch capture process was developed first and subsequently compared to continuous CaptureSMB configurations. Optimized CaptureSMB operation achieved consistent yields over four cycles, with recoveries exceeding batch operation (+14.3%) with concomitant higher productivity (+11.4%) and reduced buffer consumption (-79.2%). Critical quality attribute (CQA) analysis showed lower host cell protein levels and lower residual DNA in early CaptureSMB cycles, while full capsid ratios, thermal stability and transduction efficiency of rAAV5 particles remained unaltered across cycles and process modes. However, progressive backpressure and residual impurity accumulation in later cycles highlight operational challenges associated with direct capture of low-titer feed streams. These findings highlight that continuous twin-column CaptureSMB directly from secretion-derived perfusion harvest can improve operational efficiency, lower buffer requirements, and also maintain and, in some respects, enhance product quality. This novel strategy provides a promising route to address manufacturing capacity and cost challenges in rAAV gene therapy production.

Abstract from PubMed (PMID 42983944). This entry was added automatically by our daily literature monitor because it matches the topics we follow; read the full paper at the original source.