W
ith adeno-associated virus (AAV)-based gene transfer vector protocols recently demonstrating clear evidence of safety and long-term efficacy for some diseases including an inherited retinal disorder and hemophilia B in human clinical studies (reviewed in Mingozzi and High, 2011), the probability of clinical success for many other AAV programs at the preclinical and early clinical stages appears to have markedly increased. Critical to support this accelerating pace of clinical progress is the parallel development of vector-manufacturing capacity, that is, the ability to make sufficient quantities of safe and highly purified investigational product in accordance with current Good Manufacturing Practices (cGMP) for pivotal clinical trials and prospective licensure. As an estimate, current capacities at existing vector-manufacturing facilities range from 1014 to 1016 vector genomes (VG) (approximately 1–100 mg of purified vector) per lot, amounts adequate for preclinical and early-phase clinical studies for most disease programs, as well as late-stage clinical development and commercialization for rare diseases requiring only modest doses per patient, but insufficient to support advanced clinical development for more common diseases. It seems likely that at least a 100-fold higher manufacturing capacity, that is, cGMP lots in the range of 1016 to 1018 VG (0.1–10 g) will be required to meet the requirements of late-stage clinical development and product licensure for many recombinant AAV products, especially those aimed at the most commercially viable disease applications. Coupled with increased manufacturing capacity is the need to lower per-dose costs, and to ensure product safety using large-scale manufacturing processes. In this issue of Human Gene Therapy, Cecchini and colleagues describe a large-scale manufacturing process for recombinant AAV that demonstrates excellent potential to meet these capacity, safety, and cost-efficiency requirements. They report consistent production and recovery of >1016 VG of purified vector from 200 liters of suspension cell culture. The vector generation (upstream) methodology used is based on recombinant baculovirus infection of insect cells, first reported (Urabe et al., 2002) and since subjected to extensive process development and improvements by the Kotin laboratory (the authors of the current paper) at the National Heart, Lung, and Blood Institute (NIH, Bethesda, MD) and others. Cecchini and colleagues use cryopreserved baculovirus-infected insect cells (BIICs) (Wasilko et al., 2009) for more efficient inoculation of large-scale production insect cells, a simplified strategy for large-scale inoculation that also addresses challenges relating to the stability of multiply passaged baculovirus master and working banks. The large-scale cell culture is performed with serum-free cell culture medium, an important feature that increases safety (by eliminating a potential source of adventitious agents contamination), reduces manufacturing costs, and simplifies subsequent purification. For the downstream (purification) process, Cecchini and colleagues incorporate a series of biopharmaceutical industry-standard steps including continuous flow harvest homogenization, clarification by capsule filtration, immune-affinity and size-exclusion column chromatography, and tangential flow filtration. In addition to substantial removal of other impurities, the levels of empty capsids, a typically abundant vector-related impurity difficult to remove by scalable purification process steps, were found to represent the minor fraction of AAV particles in the final purified vector. Importantly the authors report remarkably consistent normalized final yields for their process in terms of volumetric (6.5×1013 VG/liter) and per cell (18,000 VG/cell) productivity as the process was scaled from a 10- to 200-liter bioreactor. Documentation of such consistency supports the notion that further scale-up can result in a proportional increase in manufacturing capacity. Extrapolating 100-fold to a biopharmaceutical industry 20,000-liter “large-scale” bioreactor and assuming successful scale-up of the purification steps, the manufacture of 1018 VG appears feasible; however, such theoretical projections will have to be validated by actual process scale-up. Cecchini and colleagues report an average purification yield of approximately 20%, with the most significant vector losses occurring after the filtration steps used to clarify crude cell harvest lysates; hence, there is potential for further increases in the already impressive manufacturing capacity at the scale reported by this group. Of particular note in this study is the extensive process characterization performed, including documentation of the effects of process scale-up and careful evaluation of process consistency, bioprocess-engineering strategies that will facilitate future process validation efforts. This report represents a significant step forward in the development and implementation at large scale of a scalable and cost-effective manufacturing process for AAV vectors based on the baculovirus production method for rAAV. In conjunction with parallel optimization in vector design and clinical protocols, such advances in vector manufacturing will help to ensure that the exciting momentum generated by recent successes in human gene therapy will continue.
