With the explosive growth in cell and gene therapy clinical trials, there has been considerable stress placed on manufacturing these therapies at scale. One challenge that has received attention is the production of viral vectors for gene delivery. Viral vectors are adapted from viruses that infect mammalian cells and have been adapted to grow in animal or insect cell culture systems. The vectors are used to deliver the therapeutic genes into targeted cells. Adeno-associate virus (AAV) is the most commonly used vector platform because it elicits a low inflammatory response (i.e., non-pathogenic) compared to other vectors and can be targeted to different bodily tissues. Traditionally, it has been cultured in HEK293 and HEK293T cell lines due to transfection efficiencies and a lack of expressed viral restrictions. Unfortunately, AAVs are limited in their total DNA packaging capacity (5 kilobases), which limits its application for larger gene loads. Additionally, AAV use has exceeded supply, so there is a pressing need for both an increase in vector-manufacturing capacity and an increase in AAV process productivities.

Groups at the Boston University School of Medicine and the National Defense Medical College in Japan recently described an improved method of AAV production and purification from both spent cell media and cell lysates. The team was motivated to increase vector yields and reduce contamination since most published protocols were laborious and had poor results. The study utilized HEK293T cells grown in T150 flasks under a variety of media conditions. To obtain a baseline titer of the vector, the team used a traditional media composition composed of Dulbecco Modified Eagle Medium (DMEM) containing 25 mM glucose, 2% fetal bovine serum (FBS) and 2 mM of the dipeptide supplement L-alanyl-L-glutamine. This dipeptide replaced the traditional glutamine supplementation to avoid increases in ammonia levels in the media due to glutamine decomposition. Reducing the FBS concentration in half did not reduce titer and decreased protein contamination after purification. In subsequent studies, the team added HEPES and sodium bicarbonate to improve the buffer capacity of the media. This was done in an attempt to limit the observed drop in the pH of the media due to viral protein expression, but the additions did not impact titer. The team maintained these changes and reduced the glucose concentration to 5 mM to decrease the lactate released into the cultures from glycolysis. The pH of the media stabilized to ~7.4, and the viral titers increased 3-fold, indicating that viral production is improved by limiting glycolytic acidification of the media.

The study also detailed methods of harvesting the AAVs from both media and cell lysates. The team used a simple PEG/NaCl precipitation to purify vectors from the media and a gentle cell lysis procedure with an acidic citrate buffer to decrease the contamination risk from cellular proteins and DNA. These lysates also underwent the PEG precipitation step, and all extracts from both media and lysates were extracted with chloroform. For in vitro analysis, the AAVs were used as-is. For in vivo studies, the extracts were then purified through a simple aqueous two-phases partitioning consisting of PEG (proteins and other contaminants) and ammonium sulfate (AAVs). The AAVs were then purified with a discontinuous iodixanol gradient centrifugation step to remove any remaining contaminants while separating empty or incomplete viral particles from ones to be used in vivo. With the addition of this step, the resulting vectors were very pure (>90%).

For functional tests, the purified AAVs were tested first in an in vitro study of primary mouse hepatocytes and skeletal muscle-derived C2C12 cells. The cells responded to the AAVs by decreasing the targeted gene expression by 70%. Testing of the AAVs in vivo in mice had similar positive results. First, a targeted gene was knocked down by 66% in mouse liver 2-weeks after administration. In another experiment, intramuscular injection of the AAVs transduced the target gene in 3-weeks, albeit, with some noticeable inflammation at the injection site.

Challenges will remain for viral vector manufacturing as the growth of promising cell and gene therapies continues. Fast and economical procedures to manufacture sufficient high purity vectors are needed to support both research and clinical studies. The approach outlined in this study provides a comprehensive framework for other groups to work from and modify. Most notably, the teams focused on improving overlooked aspects from upstream processing of media optimization and downstream purification from both cell lysates and spent cell media to boost both the viral titers and purity.

By, Glenn A. Harris